Response by Patel et al to Letter Regarding Article, “Perioperative Cerebral Microbleeds After Adult Cardiac Surgery”

Dear Meinel et al.,

Thank you for your letter regarding the article ‘Perioperative Cerebral Microbleeds after Adult Cardiac Surgery’. It is reassuring to hear that your recent retrospective analysis of 155 valve patients confirmed that cardiopulmonary bypass (CPB) time was a significant factor in your regression models for predicting post-operative SWI lesion counts.

Thank you for highlighting previous research describing the possibility that SWI lesions could be caused by small metallic fragments in patients undergoing mechanical valve replacement. Following careful review of our data, we can confirm that 11/49 valve surgery patients received a mechanical heart valve. Mechanical valves tended to be in younger patients (median age: 54 (IQR: 50, 60) years compared to a median age of 65 (IQR: 59, 71) years for the remainder of our cohort. New CMBs were identified in 91% (10/11) of patients fitted with mechanical valves compared to 73% (47/64) of the remaining patients, but given the small numbers involved this difference was not statistically significant (p=1.00 using Fisher’s exact test). The power of our study to detect a difference between the 11 mechanical valve patients and the remainder of our cohort was estimated at 17%. The median number of CMBs was 3 (IQR: 1, 10) for patients fitted with mechanical valves compared to 2 (IQR: 0, 4.5) CMBs for other types of surgery. Although our study is underpowered, given the younger age demographic of mechanical valve patients a trend toward both higher numbers and greater prevalence of CMBs certainly warrants further investigation. Our results definitively confirm that CMBs are common following all types of cardiac surgery. Any hypothesis that CMBs might be due solely to valve abrasion can be confidently dispelled. To distinguish the relative contribution (if any) of metal valve fragments to CMBs, either autopsy studies or a much larger sample size would be required.

Your suggestion that SWI lesions seen following surgery might interfere with anticoagulation prescription because of fear of cerebral haemorrhage should certainly be a concern to clinicians, especially as multiple CMBs post-surgery are so common. This underlines the importance of understanding the underlying biology and pathophysiology of CMBs acquired through surgery. If CPB CMBs do not share the same pathophysiological features indicative of long-term bleeding risk, then SWI lesions in patients who have undergone cardiac surgery should not be a contraindication for intravenous thrombolysis or long-term anticoagulation.

It is unclear whether mapping the size, shape and location of ‘black holes’ in SWI scans would aid in distinguishing whether ‘CMBs’ are due to chronic vascular pathology (e.g. CAA), CPB CMB lesions, or signals from embolic metallic fragments. This would be interesting to explore in future research.

CAA occurs when insoluble amyloid fibrils are deposited in the walls of leptomeningeal and cortical small arteries. This chronic arteriopathy accumulates over many years, therefore, it did not seem appropriate to diagnose ‘probable CAA’ from post-operative scans. Microbleeds seen after CPB are more likely to be due to heparinisation (anticoagulation), microemboli, inflammation, or cerebral haemodynamic changes during surgery rather than amyloid deposition. This agrees with our logistic regression findings suggesting an association between CMBs and CPB time. However, we cannot rule out the possibility that perioperative factors interact with CAA, causing microbleeds to occur in already fragile amyloid-riddled vessels due to pressure changes, combined with heparinisation, leading to vessel rupture and prolonged bleeding. Further studies of cardiac surgery patients with confirmed CAA and mild cognitive impairment would be useful in confirming any interaction; we hypothesise that patients with mild cognitive impairment might be at exceptionally high-risk of new CMBs and accelerated dementia conversion as a consequence of surgery.

Unfortunately, we did not examine our MR images for superficial siderosis, so are currently unable to report the incidence of this finding, either pre- or post-operatively. We hope to be able to perform this analysis in the near future with a view to publishing our findings. Many thanks to the authors for suggesting this additional analysis may be of interest.

We can confirm that no patients had prior CPB exposure. Our records did not note any patients with infectious endocarditis or diagnoses other than stated in the article’s supplemental material. Unfortunately, we did not record coagulation parameters during surgery. As stated in the limitations of the study, a more detailed investigation of CPB parameters and potential confounders for generation of SWIs would benefit future research.

Once again, many thanks to Meinel at al. for their interest in our research.

Kind regards,

Dr Nikil Patel, Ms Claire Hannon, Dr Emma Chung