Bringing doctors together

or

A free, private network for verified physicians

Leading voices

Dr. Philip Schauer

Bariatric Surgery · PBRC

View

Dr. Justin Hovey, FAAP, FACP, DABOM

Obesity Medicine · Noland Hospital Dothan II

View

Dr. Kimberly Gudzune

Obesity Medicine · JHU

View

Dr. Richard Dutton

General Anesthesiology · UMSOM

View

Dr. Sarah Ro

Obesity Medicine · UNC Health Care

View

Dr. Rebecca Andrews

Primary Care · UCHC

View

Dr. Jed Wolpaw

General Anesthesiology · Johns Hopkins

View

Dr. Matt Sherrer

Perioperative Med · UAB

View

Dr. David Bates

General Internal Medicine · BWH

View

Dr. Reema Dbouk

Obesity Medicine · Emory University

View
  • Discover your community

  • Stay current

  • A profile that represents you

  • Find your people

About RoonPrivacyTerms

    Overview

    Association of Quality of Life Domains and Clinical Symptoms in Patients With Familial Cerebral Cavernous Malformation.

    Aug 29, 2025DOI: 10.1161/jaha.124.035926Journal ArticleAHA

    Written by: Cynthia TsangJeffrey NelsonRachel VassarNerissa U KoCharles McCullochEdward R SmithSudhakar VadiveluAmy AkersCornelia LeeJoseph ZabramskiMichel T TorbeyIssam A AwadHelen KimBrain Vascular Malformation Consortium CCM Investigators

    Find these authors on Roon

    RV
    Dr. Rachel Vassar
    Peds Vascular Neurology · UCSF

    METHODS

    The data that support the findings of this study are available from the corresponding author upon reasonable request.The BVMC (Brain Vascular Malformation Consortium) CCM study is an ongoing prospective cohort study enrolling and following fCCM participants since 2009.22 In the first cycle (BVMC1: 2009–2014), eligible participants included those with a confirmed founder mutation in CCM1 Q455X enrolled at a single site (University of New Mexico). In the second cycle (BVMC2: 2014–2019), eligible participants had any genetic mutation in CCM1 or met 2 of 3 of the following clinical criteria: (1) clinical diagnosis of CCM, (2) multiple CCMs on magnetic resonance imaging, or (3) family history of CCM. BVMC2 participants were enrolled at 4 recruitment sites (University of New Mexico, University of California San Francisco, Barrow Neurological Institute, and Alliance to Cure Cavernous Malformation). In the third cycle (BVMC3: 2019‐2025), eligible participants had either: (1) confirmed genetic mutation in CCM1, CCM2, or CCM3; or (2) individuals who met 2 of the 3 clinical criteria as described above for BVMC2. BVMC3 participants were recruited from 6 enrollment sites: Barrow Neurological Institute, Boston Children's Hospital, Cincinnati Children's Hospital, University of Chicago, University of New Mexico, and University of California San Francisco. Local institutional review board approval was obtained for the study at each participating site in BVMC1 and BVMC2. All participating sites relied on University of California San Francisco as the single institutional review board for BVMC3. Written informed consent or assent, as appropriate, was obtained for all study participants.This study was a cross‐sectional analysis of enrolled fCCM adults (≥18 years old) in the BVMC study who completed a PROMIS‐29 questionnaire at either baseline enrollment or follow‐up visit between 2019 and 2022.22At enrollment and annual follow‐up visits, clinical and outcome data were collected. CCM clinical symptoms (eg, history of ICH, seizures, headaches, focal neurological deficits), treatments, and events were abstracted from medical records or from a patient interview using standardized case report forms. Functional outcome of patients was assessed using the mRS score, which is a simple 6‐point scale reflecting increasing disability and ranging from 0 (no symptoms) to 6 (death).16 In the BVMC3 cohort, patient‐reported QoL over the past 7 days was assessed using PROMIS surveys. For adults, PROMIS‐29 version 2.0 short forms assessed 7 health domains (physical function, fatigue, pain interference, depressive symptoms, anxiety, ability to participate in social roles and activities, and sleep disturbance) using 4 items per domain, with responses based on a 5‐point Likert scale.23 The PROMIS‐29 survey (hereafter simply referred to as PROMIS) was completed by patients through phone, mail, in person, or online. If questionnaires were completed by phone and in some cases in person, questions were read aloud to patients by study coordinators, and patient responses were recorded with assistance by a surrogate in rare cases if needed, for example, if speech was impaired. All study data were collected and managed using Research Electronic Data Capture tools hosted at the Data Management and Coordinating Center at the University of Cincinnati.24, 25Descriptive statistics including frequencies, proportions, mean±SD, and median and interquartile range (IQR) are reported for patient characteristics. For this cross‐sectional analysis, the earliest PROMIS questionnaire and corresponding clinical data at the time of PROMIS assessment were used. Raw PROMIS scores for each domain (ranging from 4 to 20 points) were converted to T scores, standardized to a general US reference population (mean 50±10) and oriented so that higher scores reflect worse QoL on that domain.24, 25 Intercept‐only, linear mixed models were used to compare whether mean T scores for each domain differed significantly from 50 (P≤0.05). To accommodate within‐family correlation, we included family as a random intercept. We used the Kenward‐Roger approximation to obtain more accurate inference for fixed effects.26 The intercept‐only regression model is comparable with a 1‐sample t test. Thus, someone with a mean score of 70 on a PROMIS domain is interpreted as 2 SDs above the reference mean, whereas a score of 40 would be 1 SD below the mean. Internal consistency (reliability) of PROMIS measures was assessed using the Cronbach α coefficient; values ≥0.8 are considered good and ≥0.9 as excellent.Linear mixed models were used to test whether PROMIS domain T scores (outcomes) were associated with CCM symptoms (predictors), including history of ICH, seizures, or headaches, age, and sex. We included interaction terms in the model to evaluate differences in CCM symptoms by age or sex. Using guidelines for interpreting standardized effect sizes, a difference between groups of 2 points (0.2 of an SD) indicates a small effect size, 5 points (0.5 of an SD) indicates a medium effect size, and 8 points (0.8 of an SD) indicates a large effect size.27Construct validity was evaluated through examination of score distributions and floor and ceiling effects, and by examining R 2 values from linear mixed models comparing PROMIS domain scores (outcome) and mRS (predictor) scores, reflecting the degree of variance explained by mRS. R 2 values were calculated with the Stata package R2_MLM, which is based on the methodology of Rights and Sherba and accounts for both the predictor and cluster‐specific intercepts.28, 29 We generated box plots and forest plots of PROMIS domain T‐score results. If a participant had a perfect score for a domain, then these scores were randomly changed by small amounts for better visual clarity on the box plots. All data analyses were performed and figures generated using StataNow 18.5 SE.30The data that support the findings of this study are available from the corresponding author upon reasonable request.

    RESULTS

    One hundred ninety‐eight adult fCCM participants completed at least 1 PROMIS questionnaire; summary statistics of the cohort are presented in Table 1. The median age was 49.88 years (IQR, 35.36–61.08). The sample was predominantly women (67.17%) and White race (94.24%). Over half of participants self‐identified as Hispanic ethnicity (54.45%). Most individuals with genotype information had a mutation in CCM1 (85/107, 79.44%), but only 54.04% of the cohort had available data on CCM gene mutation. At the time of PROMIS assessment, 80 participants (40.40%) had experienced at least 1 symptomatic ICH, 77 (38.89%) had a history of seizures, and 137 (69.19%) had a history of headaches.Characteristics of 198 Adult Participants With Familial CCM Enrolled in the Brain Vascular Malformation Consortium CCM Study Between 2019 and 2022CCM indicates cerebral cavernous malformation.Summary values are presented as median (interquartile range), n (percent), or n/nonmissing (percent).Distribution of PROMIS T scores in patients with fCCM are shown in Figure 1A, whereas differences from the reference population are shown in Figure 1B. The fCCM group reported worse anxiety scores (52.71 [95% CI, 51.23–54.20]; P<0.001), worse pain scores (52.50 [95% CI, 50.93–54.07]; P=0.001), and worse physical functioning scores (52.96 [95% CI, 51.40–54.52]; P<0.001) than the reference population. However, they reported better social participation scores (46.95 [95% CI, 45.35–48.54]; P<0.001). Other domains were not significantly different. We observed high reliability of items, that is, internal consistency, for each of the PROMIS domains (Cronbach α >0.85) in our sample of adult patients with fCCM (Table S1).A, Boxes show 25th, 50th, and 75th percentiles of scores. B, Domain scores to the right of 0 indicate worse and to the left of 0 indicate better scores than the reference population (vertical dotted line centered at 0). Circles represent mean differences from 50 and horizontal lines indicate 95% CIs. CCM indicates cerebral cavernous malformation; and PROMIS, Patient‐Reported Outcomes Measurement Information System.Next, we looked at associations between age, sex, and clinical symptoms with PROMIS domain scores. Older age was associated with worse depression (+0.87 points per decade increase in age [95% CI, 0.12–1.61]; P=0.023), worse pain (+1.34 per decade [95% CI, 0.47–2.21]; P=0.003), worse physical functioning (+1.90 per decade [95% CI, 1.04–2.75]; P<0.001), and worse social satisfaction (+1.66 per decade [95% CI, 0.79–2.53]; P<0.001). Women had worse sleep disturbance scores compared with men (+4.28 [95% CI, 1.54–7.02]; P=0.002). No significant interactions were observed between age or sex and history of ICH, seizures, or headaches for any of the PROMIS domains (data not shown). Recruitment site was considered a potential confounder but was not associated with domains, did not improve the goodness of fit of the model, and did not appreciably change results, so it was not included in the model (data not shown).We then assessed differences in PROMIS scores of patients with fCCM with and without a history of ICH, seizures, or headaches (Figure 2) at the time of PROMIS assessment, with multivariable regression analyses adjusting for age and sex (Table 2). We found that participants with prior ICH had worse anxiety (+3.11 [95% CI, 0.10–6.12]; P=0.043), fatigue (+3.68 [95% CI, 0.61–6.76]; P=0.019), physical function (+3.79 [95% CI, 0.74–6.84]; P=0.015), and social participation scores (+4.42 [95% CI, 1.28–7.56]; P=0.006) compared with those without prior ICH. History of seizures was not associated with any PROMIS domain. A history of headaches was associated with worse depression (+3.17 [95% CI, 0.18–6.17]; P=0.038), fatigue (+4.14 [95% CI, 0.70–7.59]; P=0.019), pain (+3.88 [95% CI, 0.40–7.37]; P=0.029), and sleep disturbance scores (+2.94 [95% CI, 0.03–5.85]; P=0.048). Fatigue was the only domain significantly associated with >1 symptom (prior ICH and history of headaches). Anxiety was also associated with worse scores in those with versus without prior ICH and headaches (>3.0 points higher), although it did not reach statistical significance for headaches (P=0.064).A positive value indicates a worse score for those with the symptom. CCM indicates cerebral cavernous malformation; and PROMIS, Patient‐Reported Outcomes Measurement Information System.Association of PROMIS‐29 Domains and Clinical Symptoms in Participants With Familial Cerebral Cavernous MalformationA positive effect size indicates an unfavorable score for the group with the symptom. PROMIS indicates Patient‐Reported Outcomes Measurement Information System.Effect sizes that are moderate in size according to Cohen's criterion (>0.3 SDs).Regression coefficients that are statistically significantly different from 0 at P<0.05.Finally, we examined the correlation and variance explained between PROMIS and mRS scores (Table 3). Significant positive correlations were observed between mRS score and all PROMIS domains except for sleep disturbance (P=0.124). Strongest associations were observed for physical function (+6.40 points per mRS score increase [95% CI, 5.42–7.38]; P<0.001), social participation (+5.25 [95% CI, 4.16–6.35]; P<0.001), and pain scores (+3.15 [95% CI, 1.88–4.42]; P<0.001). In general, as mRS scores increased, so did the mean PROMIS domain scores (Figure S1). However, the variance explained by mRS score was moderate to weak (R 2 of 0.55 for social participation, 0.47 for physical function, 0.16 for pain, and <0.10 for all other PROMIS domains). Thus, PROMIS provides additional valuable information beyond the mRS.Association of PROMIS Health Domains With Modified Rankin Scale ScorePROMIS indicates Patient‐Reported Outcomes Measurement Information System.

    DISCUSSION

    In the largest study of QoL in fCCM populations to date, we assessed QoL domains through PROMIS surveys and evaluated the impact of clinical symptoms on domain scores. We found that patients with fCCM reported significantly different responses in anxiety, pain, physical functioning, and social participation compared with the general US reference population. Participants with a prior history of ICH reported worse scores for anxiety, fatigue, and physical functioning compared with those without ICH. Those with chronic headaches reported significantly worse scores for depression, fatigue, pain, and sleep disturbance compared with those without headaches. There were no significant differences observed for any domains in those with and without seizures. Fatigue was the only PROMIS domain significantly associated with >1 clinical symptom (prior ICH and history of headaches).Given the heterogeneity of clinical presentations even among fCCM cases, our study provides insight into disease morbidity and associations with common symptoms of the disease. There is no CCM‐specific QoL survey available yet; thus, many studies have used similar or different generic patient‐reported metrics for assessing QoL. Three other CCM studies have used PROMIS QoL surveys. Among 110 patients with CCM with symptomatic ICH within the past year who were enrolled in a prospective trial‐readiness project, nearly half reported PROMIS scores at baseline that differed significantly from the general reference population (eg, less depressed, more socially satisfied, but more anxious).20 Thirty percent reported at least 1 affected domain at 2 years follow‐up with improvement in social function, anxiety, and pain.21 In a study of 104 patients with brainstem CCM, 64% had at least 1 PROMIS domain ≥1 SD from the general population, including 35% with increased fatigue, 36% with increased anxiety, and 28% with greater social participation scores.17 Other studies have also reported worse health‐related domains in patients with CCM compared with reference populations assessed using the Short‐Form 36 Health Survey,18, 19, 31, 32 the Hospital Anxiety and Depression Scale (versions A and D),19, 31 or EuroQoL.18 Compared with age‐ and sex‐matched controls, 219 patients with CCM showed significantly decreased scores, indicating worse health status, in all 8 domains of the Short‐Form 36 Health Survey (general health, vitality, physical ability, and limitations in social activities due to physical and social problems).31 Thirty‐nine percent of patients showed increased anxiety on the Hospital Anxiety and Depression Scale compared with 20% observed in a normal population.31 These results were mirrored in the subset of 47 patients with CCM‐related with epilepsy.19 In 49 untreated patients with CCM, investigators observed worse health outcomes through decreased Karnofsky Performance Status survey scores in pain, vitality, and general health perception fields.18 In contrast, 1 recent study in 71 patients with fCCM participating in the phase I/II propranolol for familial cerebral cavernous malformation (TreatCCM) trial conducted in Italy found no significant differences in anxiety and physical and mental subscales of the Short‐Form 36 Health Survey, but observed significantly lower rates of depression in treated versus nontreated groups after 2 years using the Beck Depression Inventory II scale.32 In our study, we also observed that many of these same domains were associated with worse scores in patients with fCCM with prior ICH or history of headaches, but interestingly not with a history of seizures. Similarly, there was no significant effect of seizures on depression, anxiety, and mental or physical Short‐Form 36 Health Survey subscales observed in the TreatCCM trial either.32 Differences may be due to different comparison groups (general US versus Italian populations, CCM cases without clinical symptoms, or on therapy) or timing or severity of seizure events in relation to QoL assessments (hospital versus outpatient‐based). Despite different survey methodologies, anxiety, fatigue, and physical function are commonly reported domains affected across different CCM populations, including those with fCCM.These results have important implications when considering counseling approaches and care options for patients with CCM. In addition to physical limitations, recognizing that anxiety is a consistently impacted QoL domain across multiple measurement tools in CCM populations warrants greater attention and resources toward evaluating and addressing patients' unease and worries. This is especially worthy of emphasis, because anxiety is understood to have strong association with worse health outcomes in many different populations, including in patients with cancer,33, 34 asthma,35 cardiovascular disease,36, 37 and musculoskeletal disease.38, 39 Future studies can explore the relationship of anxiety and various other mental components of disease with outcome and perception of health. If a strong relationship is shown, screening and interventions centered on anxiety reduction will be critical to improving outcomes.Finally, hypoxia or lack of oxygen has been shown to exacerbate CCM lesion formation in mouse models by promoting angiogenesis, neuroinflammation, and vascular thrombosis.40, 41 As such, patients with sleep apnea (often underdiagnosed), resulting in varying levels of hypoxia, may be at increased risk of CCM symptom exacerbation. This could explain why fatigue and sleep disturbance domain scores were reportedly worse in patients with fCCM with symptoms versus no symptoms.To our knowledge, no prior studies have evaluated psychometric properties of the PROMIS tool in fCCM. We found good to excellent reliability of PROMIS items ranging from 0.85 to 0.97 in our sample of adult patients with fCCM. These results are similar to those reported in other populations using PROMIS, such as in Parkinson disease (all Cronbach α >0.82)42 and ankylosing spondylitis (all >0.90),43 suggesting that PROMIS measures are reliable for assessing health domains in a variety of conditions, including fCCM. In addition, we found good construct validity between mRS and the PROMIS social participation and physical functioning domains, with R 2 of 0.55 and 0.47, respectively. Although mRS was significantly associated with other PROMIS domains, correlations were weaker, and no association was observed with sleep disturbance. These results suggest that PROMIS measures have added value beyond mRS as an outcome metric.Our study has several strengths and limitations. This study is the first to analyze and evaluate QoL using PROMIS surveys in a large fCCM population recruited from multiple Alliance to Cure Cavernous Malformation Centers of Excellence sites and is likely generalizable to the majority of patients with fCCM seen at other large referral centers. In addition, all sites used similar recruitment and data collection strategies; thus, we would not expect associations with symptoms to be site‐dependent, and site was not a confounder of results. Another strength is that we used a validated general health‐related QoL survey that was short, with multiple modes of administration (eg, in‐person, phone, mail, or online), thereby improving patient compliance for completion. Because the analysis only included adults and primarily consisted of CCM1 participants, results may not be generalizable to pediatric patients, CCM2/CCM3 mutation patients, or patients with sporadic CCM. A limitation of PROMIS and other patient‐reported outcomes is that patients with severe neurological deficits may be more likely to have missing assessments. However, PROMIS scores showed more variability and were not well explained by provider‐reported mRS scores even for the physical functioning domain, revealing more subtle changes and additional impaired domains in patients with fCCM that will be important to further explore.Future studies should examine longitudinal changes in patient‐reported scores, consider how QoL domains are affected in pediatric patients diagnosed with fCCM, or focus on how patient‐reported scores differ by CCM genotype or other measures of CCM disease severity. It may also be informative to consider how QoL metrics are affected by timing of survey completion in relation to timing of ICH or frequency of seizures or headaches. Further optimization, psychometric evaluation, and standardization of questionnaires for use in fCCM populations will be important for future clinical trials. Reliable and valid measures of patient‐reported outcomes that correlate with clinical symptoms in patients with fCCM, such as PROMIS, have the potential to inform clinical trials and assist clinicians in prioritizing treatments and therapies that address symptoms with the greatest morbidity in patients with CCM.

    APPENDIX

    Amy Akers, PhD; Issam A. Awad, MD; Blaine L. Hart, MD; Helen Kim, PhD; Michael T. Lawton, MD; Cornelia Lee, PsyD; Marc C. Mabray, MD; Douglas A. Marchuk, PhD; Charles E. McCulloch, PhD; Leslie Morrison, MD; Ludmila Pawlikowska, PhD; Edward R. Smith, MD; Michel Torbey, MD; Sudhakar Vadivelu, DO; Shantel Weinsheimer, PhD; Joseph M. Zabramski, MD; and Atif Zafar, MD.

    Sources of Funding

    The BVMC is part of the Rare Diseases Clinical Research Network and is funded by the National Center for Advancing Translational Sciences and the National Institute for Neurological Disorders and Stroke (U54NS065705). All Rare Diseases Clinical Research Network consortia are supported by the network's Data Management and Coordinating Center (U2CTR002818). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

    Disclosures

    Dr Kim is a consultant for Neurelis, Recursion Pharmaceuticals, and Ovid Therapeutics. She serves on the Data Safety and Monitoring Board for Neurelis. Dr Vadivelu is a consultant for Forge Biotherapeutics and a scientific advisory board member for Anuncia Medical and Longeviti. He serves on the Data Safety and Monitoring Board for Alcyone Therapeutics. Dr Awad is a consultant for Neurelis and Ovid Therapeutics.

    Supporting information

    Table S1Figure S1
    RVDr. Rachel Vassar

    Dr. Rachel Vassar

    Pediatric Neurology · Pediatric Vascular Neurology
    University of California, San Francisco
    rachelvassar