Cerebral microbleeds and stroke risk after ischaemic stroke or transient ischaemic attack: a pooled analysis of individual patient data from cohort studies

Duncan Wilson, Gareth Ambler, Keon‐Joo Lee, Jae‐Sung Lim, Masayuki Shiozawa, Masatoshi Koga, Linxin Li, Caroline Lovelock, Hugues Chabriat, Michael G. Hennerici, Yuen Kwun Wong, Henry Ma, Luís Prats‐Sánchez, Alejandro Martínez‐Domeño, Shigeru Inamura, Kazuhisa Yoshifuji, Ethem Murat Arsava, Solveig Horstmann, Jan Purrucker, Bonnie Lam, Adrian Wong, Young Dae Kim, Tae-Jin Song, Maarten Schrooten, Robin Lemmens, Sebastian Eppinger, Thomas Gattringer, Ender Uysal, Zeynep Tanrıverdi, Natan M. Bornstein, Einor Ben Assayag, Hen Hallevi, Jun Tanaka, Hideo Hara, Shelagh B. Coutts, Lisa Hert, Alexandros A. Polymeris, David Seiffge, Philippe Lyrer, Ale Algra, Jaap Kappelle, Rustam Al‐Shahi Salman, Hans Rolf Jäger, Gregory Y.h. Lip, Heinrich P. Mattle, Leonidas D Panos, Jean-Louis Mas, Laurence Legrand, Christopher Karayiannis, Thanh G. Phan, Sarah Gunkel, Nicolas Christ, Jill Abrigo, Thomas Leung, Chiu‐Wing Winnie Chu, Francesca M. Chappell, Stephen Makin, Derek Hayden, David Williams, M. Eline Kooi, Dianne H.k. van Dam-Nolen, Carmen Barbato, Simone Browning, Beom Joon Kim, Anil M. Tuladhar, Noortje A.m. Maaijwee, Anne Cristine Guevarra, Chathuri Yatawara, Anne-Marie Mendyk, Christine Delmaire, Sebastian Köhler, Robert van Oostenbrugge, Ying Zhou, Chao Xu, Saima Hilal, Bibek Gyanwali, Christopher Chen, Min Lou, Julie Staals, Régis Bordet, Nagaendran Kandiah, Frank-Erik de Leeuw, Robert Simister, Aad van der Lugt, John Ly, Joanna M. Wardlaw, Yannie Soo, Felix Fluri, Velandai Srikanth, D. Calvet, Simon Jung, Vincent I.h. Kwa, Stefan T. Engelter, Nils Peters, Eric E. Smith, Yusuke Yakushiji, Dilek Neci̇oğlu Örken, Franz Fazekas, Vincent Thijs, Ji Hoe HeoThe Lancet Neurology 2019 · DOI

TL;DR

This large-scale pooled analysis of individual patient data from 38 cohorts reveals that while cerebral microbleeds (CMBs) are associated with a higher relative risk of future intracranial hemorrhage (ICH) compared to recurrent ischemic stroke (IS), the absolute risk of IS remains substantially higher than ICH, regardless of CMB burden or distribution. This finding is crucial for guiding antithrombotic therapy decisions in stroke survivors.

What Problem Does This Paper Address?

A critical challenge in managing patients who have experienced an ischemic stroke (IS) or transient ischemic attack (TIA) is balancing the risk of recurrent IS against the risk of intracranial hemorrhage (ICH) when prescribing antithrombotic therapies (like antiplatelets or anticoagulants). Cerebral microbleeds (CMBs), small lesions visible on MRI, are known neuroimaging markers of small vessel disease and have been linked to both IS and ICH. However, it remained unclear whether a high burden or specific patterns of CMBs would specifically identify patients at a higher absolute risk of ICH than IS, which is essential for making informed treatment decisions. Previous studies were limited by small sample sizes, hindering definitive conclusions.

Methods

The researchers conducted a pooled analysis of individual patient data from 38 prospective cohort studies, including over 20,000 patients. They used a standardized approach to collect and analyze data on patient demographics, risk factors, MRI findings (specifically CMBs), and subsequent stroke events (both IS and ICH). Statistical analysis involved multivariable Cox regression models, adjusted for numerous prognostic and confounding variables, to estimate hazard ratios (aHRs) for the primary outcomes: a composite of symptomatic ICH or IS, symptomatic ICH alone, and symptomatic IS alone. They also performed sensitivity analyses, including a two-stage meta-analysis, to confirm the robustness of their findings.

Datasets

The study utilized data from 38 existing cohort studies, pooling individual patient data. No new datasets were created or introduced. The total dataset comprised 20,322 participants with over 35,225 patient-years of follow-up.

Key Insights and Ideas

  • Absolute vs. Relative Risk: The most significant insight is the distinction between relative and absolute risk. While CMBs increase the relative hazard of ICH, the absolute risk of IS remains higher. This is a crucial nuance for clinical decision-making.
  • CMBs as a Marker of Small Vessel Disease: The findings suggest CMBs are a marker for underlying small vessel disease that predisposes to both IS and ICH, rather than solely a predictor of bleeding.
  • No Clear Benefit of Withholding Antithrombotics: The data does not support withholding antithrombotic treatment solely based on the presence of CMBs, as the risk of IS is consistently higher.
  • Lobar CMBs and CAA: The study found no significant association between lobar CMBs (often indicative of cerebral amyloid angiopathy, CAA) and increased ICH or IS risk, suggesting potential limitations in using CMBs alone for CAA diagnosis in certain contexts.

Key Results

  • Composite Outcome: Patients with any CMBs had a 35% higher risk of the composite outcome (ICH or IS) compared to those without (aHR 1.35).
  • ICH Risk: CMBs were associated with a 2.45-fold increased risk of symptomatic ICH. This risk increased significantly with higher CMB burdens (e.g., ≥20 CMBs had an aHR of 8.61 for ICH).
  • IS Risk: CMBs were associated with a 23% increased risk of symptomatic IS (aHR 1.23). This increase was less pronounced than for ICH, even with higher CMB burdens.
  • Absolute Risk Dominance of IS: Crucially, even in patients with ten or more CMBs, the rate of IS (64 per 1000 patient-years) significantly exceeded the rate of ICH (27 per 1000 patient-years). This pattern held true for ≥20 CMBs as well.
  • Antithrombotic Therapy: No significant interactions were found between CMBs and antiplatelet or oral anticoagulant therapy for ICH risk, suggesting these treatments can generally be used safely.

Limitations and Future Work

The authors acknowledge several limitations:

  • Observational Design: The study is observational, meaning it cannot definitively prove causation. Potential for selection bias and confounding by indication (e.g., physicians prescribing antithrombotics to healthier patients) exists.
  • Event Adjudication: Not all included studies formally adjudicated all events, which could introduce variability.
  • MRI Sequence Variability: While adjusted for, different MRI sequences used to detect CMBs have varying sensitivities.
  • Rare High CMB Burdens: The study could not precisely estimate risks for extremely high CMB burdens (e.g., ≥50) due to their rarity.
  • Residual Confounding: Despite extensive adjustments, unknown or uncontrolled factors could still influence the results.
  • Net Clinical Benefit: The study did not formally assess net clinical benefit, which would weigh the severity of ICH against recurrent IS.

Future work suggested includes:

  • Randomized Controlled Trials (RCTs): To definitively establish whether CMBs modify the net clinical benefit of antithrombotic drugs, RCTs are needed.
  • Large Sample Sizes for Interactions: Identifying statistically significant interactions between CMBs and antithrombotic therapies will likely require very large sample sizes, as seen in ongoing trials.
  • Biomarker Integration: Further research is needed to explore the utility of neuroimaging biomarkers like CMBs in improving risk prediction scores for both ICH and IS.

Summary updated

Abstract

British Heart Foundation and UK Stroke Association.
BibTeX
@inproceedings{wilson2019cerebral,
  author = {Duncan Wilson and Gareth Ambler and Keon‐Joo Lee and Jae‐Sung Lim and Masayuki Shiozawa and Masatoshi Koga and Linxin Li and Caroline Lovelock and Hugues Chabriat and Michael G. Hennerici and Yuen Kwun Wong and Henry Ma and Luís Prats‐Sánchez and Alejandro Martínez‐Domeño and Shigeru Inamura and Kazuhisa Yoshifuji and Ethem Murat Arsava and Solveig Horstmann and Jan Purrucker and Bonnie Lam and Adrian Wong and Young Dae Kim and Tae-Jin Song and Maarten Schrooten and Robin Lemmens and Sebastian Eppinger and Thomas Gattringer and Ender Uysal and Zeynep Tanrıverdi and Natan M. Bornstein and Einor Ben Assayag and Hen Hallevi and Jun Tanaka and Hideo Hara and Shelagh B. Coutts and Lisa Hert and Alexandros A. Polymeris and David Seiffge and Philippe Lyrer and Ale Algra and Jaap Kappelle and Rustam Al‐Shahi Salman and Hans Rolf Jäger and Gregory Y.H. Lip and Heinrich P. Mattle and Leonidas D Panos and Jean-Louis Mas and Laurence Legrand and Christopher Karayiannis and Thanh G. Phan and Sarah Gunkel and Nicolas Christ and Jill Abrigo and Thomas Leung and Chiu‐Wing Winnie Chu and Francesca M. Chappell and Stephen Makin and Derek Hayden and David Williams and M. Eline Kooi and Dianne H.K. van Dam-Nolen and Carmen Barbato and Simone Browning and Beom Joon Kim and Anil M. Tuladhar and Noortje A.M. Maaijwee and Anne Cristine Guevarra and Chathuri Yatawara and Anne-Marie Mendyk and Christine Delmaire and Sebastian Köhler and Robert van Oostenbrugge and Ying Zhou and Chao Xu and Saima Hilal and Bibek Gyanwali and Christopher Chen and Min Lou and Julie Staals and Régis Bordet and Nagaendran Kandiah and Frank-Erik de Leeuw and Robert Simister and Aad van der Lugt and John Ly and Joanna M. Wardlaw and Yannie Soo and Felix Fluri and Velandai Srikanth and D. Calvet and Simon Jung and Vincent I.H. Kwa and Stefan T. Engelter and Nils Peters and Eric E. Smith and Yusuke Yakushiji and Dilek Neci̇oğlu Örken and Franz Fazekas and Vincent Thijs and Ji Hoe Heo},
  title = {{Cerebral microbleeds and stroke risk after ischaemic stroke or transient ischaemic attack: a pooled analysis of individual patient data from cohort studies}},
  booktitle = {The Lancet Neurology},
  year = {2019},
  doi = {10.1016/s1474-4422(19)30197-8},
}
Figure