Article Text

Download PDFPDF

Risk and resiliency: thrombotic and ischemic vascular events, in cyanotic congenital heart disease
  1. Craig S Broberg
  1. Correspondence to Dr Craig S Broberg, Adult Congenital Heart Program, Department of Cardiovascular Medicine and Radiology, Knight Cardiovascular Institute, Oregon Health & Science University, Portland, Oregon 97239, USA; brobergc{at}

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Adults with cyanotic congenital heart disease, including Eisenmenger syndrome, are a remarkable group of people. While there is an eagerness to cite the historic increase in survival of patients with congenital heart diseases, outside these successes are those for whom surgery either was not or could not be offered who, nevertheless, beat the odds and survived. These cyanotic individuals often relate stories of being told they would not live to reach the age of 4 or 12 or 20, yet, in an extraordinary show of resilience, are now in their fifth decade or beyond. For most, exertional tolerance is below normal, but they, otherwise, carry on with life, including careers and family. They tell anecdotes of healthcare providers who, upon noting their low oxygen saturation, erroneously raise an alarm for immediate resuscitation. They defy providers’ expectations by maintaining full-time employment, flying abroad, living at high altitude1 or surviving neurosurgery. They have important physiological adaptations that allow them to carry on, although with an ongoing risk that their fragile balance could be unfavourably tipped at any time.

Recognising both their remarkable survivorship and ongoing risks, cyanotic adults with congenital defects merit focused scientific consideration despite their low numbers. It is, therefore, gratifying to see researchers and editors give attention to these unique survivors. Investigators from Copenhagen have pooled data on over 100 cyanotic adults from Sweden and Denmark for several descriptive investigations.2 ,3 In their Heart paper, the Danes explore the prevalence of cerebral and pulmonary thromboembolic ‘events’ through comprehensive imaging.4 They studied 98 patients with brain MRI, CT pulmonary angiography and/or ventilation/perfusion (V/Q) scintigraphy or single-photon emission CT (SPECT). The majority had Eisenmenger physiology. Twenty per cent had a clinical history of thrombotic events (requiring hospitalisation) whereas imaging evidence of either cerebral or pulmonary phenomena was present in 47% and 31%, respectively, demonstrating an importantly high prevalence of such findings. The study excludes some who may have previously died of related vascular complications such as haemoptysis, cerebral abscess or coronary emboli through similar mechanisms. Thus, the prevalence figures described herein may be underestimates.

Given that vascular phenomena have been previously described, the findings seem more confirmatory than revelatory, and the higher prevalence may reflect more sensitive imaging protocols. However, the study supports two other substantive conclusions. First, there are discrepancies between events detected by imaging versus clinical history. Second, different mechanisms of events, both thrombosis and ischaemia, occur with different frequencies between the pulmonary and cerebral circulation.

Pulmonary thrombosis

Pulmonary vascular phenomena were diagnosed either directly by CT (20% positive) or indirectly by V/Q mismatch (29% positive). Previous Eisenmenger studies show similar prevalence of intravascular thrombus by CT, including layering mural thrombi in large arteries,5 ,6 but comparisons between V/Q and CT findings are new.

The data are consistent with the notion that vascular arteriosclerotic changes and pulmonary artery enlargement may lead to thrombosis in situ, and that thrombotic fragments may break from the thrombus and flow peripherally to lodge in the microcirculation.6 This has been a proposed mechanism for haemoptysis,7 and was present in Eisenmenger's first index case. In the Danish study, pulmonary vascular calcifications were present mostly in patients with Eisenmenger syndrome, consistent with this mechanism in pulmonary hypertension. Yet, only six participants with imaging findings had a clinical history of pulmonary thrombosis. Furthermore, not every affected individual had abnormalities on both tests. Of 21 participants with any evidence of pulmonary thrombosis, four had positive CT scans only. Seven demonstrated V/Q mismatching without CT findings, indicating alternative embolic sources.

Cerebral infarction

Their data suggest that while the pulmonary circulation may be affected by atherosclerotic changes from pulmonary hypertension, the cerebral circulation is not. Conversely, the cerebral circulation is more vulnerable to low oxygen tension; the pulmonary circulation is less so. Out of 72 individuals studied with brain MRI, strikingly, 47% of patients had findings of a prior stroke, and over half of these had more than one (29% had both lacunar and cortical lesions), often in areas prone to ischaemia. The plausible presumption in this context is that each defect represents a prior episode of low oxygen tension long enough to induce tissue changes. Cerebral findings were more prevalent on imaging than expected from clinical history alone, so, the process occurs more frequently than previously believed. Though unproven, the authors justifiably surmise that a portion of these cerebral events may have been induced by healthcare providers either through phlebotomy or microemboli at the time of venous access. If nothing more, this serves as a reminder to clinicians of the need for caution when considering any intervention, and to obligatorily use filters during venous infusion. There was a trend towards fewer events in those with iron deficiency (OR 0.39, range 0.14–1.08), despite mixed literature suggesting microcytosis is related to stroke. Iron deficiency does not increase viscosity however,8 consistent with the data here indirectly suggesting that microcytosis may actually be protective.

The data also found that 5 out of 12 patients with a clinical neurologic event previously had nothing to show for it on imaging. Presumably, these reflect temporary ischaemia from transient poor cerebral oxygen transport. Cerebral findings were also associated with lower resting oxygen saturation, a marker in many studies of disease severity despite its dynamic nature. Possible triggers for cerebral ischaemia or infarction might include suboptimal haemoglobin (from iron deficiency, phlebotomy, significant epistaxis or haemoptysis), supraoptimal haemoglobin (causing flow-limiting hyperviscosity), hypermetabolic states (such as infection) or poor systemic blood flow (from transient arrhythmia or ventricular dysfunction). Notably, 24% had atrial arrhythmia.

Findings considered

Clearly, cyanotic individuals are prone to vascular injuries, but not all develop them. More than half were spared. Why? What mechanisms allow so many to remain free of trouble despite oxygen saturations that would render most of us unconscious? We speculate that events stem from too high or too low a haemoglobin or haematocrit (the former being the determinant of oxygen delivery, the latter being the determinant of viscosity),8 Yet, since over half the population are spared, they must reach an optimal compensation of oxygen delivery via mechanisms with which we should probably not interfere. The authors found no difference in haematocrit (and presumably haemoglobin) to discriminate affected patients, or differences in platelet count, iron levels or thromboelastography. This was a one-time cross-sectional study, whereas erythropoiesis is a dynamic process driven in response to need. It could be that vascular events occurred during previous times of either too high or too low a blood count. Still, in the authors’ view, it is one more reminder that an elevated haematocrit in and of itself should not be viewed as a problem; that erythropoiesis should foremost be considered adaptive, not maladaptive, despite some of its inherent negative consequences.3

There are acknowledged limitations in the study. Not every patient received every investigation, an expected reality in clinical research of this kind, which leaves some interpretive holes. Yet, the cohort of those that were comprehensively studied (N=57) is still a respectable sample size of a condition that in Western countries is slowly receding. Imaging studies do not determine the time of events, which would require longitudinal study; something no investigation has performed. Still, the message from these findings is consistent with previous research; patients are vulnerable to vascular complications, both ischaemic and embolic, even more than previously believed.

The study does not answer some haematological questions that have lingered from relevant work in the past, such as how to strike the right balance between oxygen delivery and hyperviscosity,8 the relationship between haemoptysis and pulmonary thrombosis and the role, if any, of anticoagulation or phlebotomy.1 A substantial portion of patients with Eisenmenger syndrome experience haemoptysis (all patients by age 40 in Paul Wood's seminal series),7 and many die from it. It would be of interest to understand the relationship, if any, between V/Q mismatching and prior haemoptysis.

Another important question is the potential role of anticoagulation. Ascribing to the hypothesis that haemoptysis occurs secondary to microthrombi in the distal pulmonary vasculature, Wood advised that treatment should focus on ‘preventing secondary thrombo-obstructive lesions by means of permanent anticoagulants, such as phenindione’.7 In light of the common prevalence found by the authors in this study, it is tempting to believe such a strategy could be beneficial. However, the authors found no difference in those on anticoagulation, consistent with prior work,3 although retrospective. Anticoagulation is not benign in these patients for a number of reasons, haemoptysis is one among them.

Despite the remaining questions, the present study is a provocative reminder that successful survivors of cyanotic heart disease are still susceptible to both thrombosis and ischaemia, that physiological adaptations to cyanosis are far more complicated than our current understanding of them, that our interpretation of clinical signs and symptoms is imperfect and that extrapolation of our own views about the role of treatments such as phlebotomy and anticoagulation remain uninformed. Thoughtful studies like this in one of the most intriguing, but often forgotten, patient groups in cardiology continue to offer insights into both the susceptibility and resiliency of the patient with cyanosis.


View Abstract


  • Competing interests None declared.

  • Provenance and peer review Commissioned; internally peer reviewed.

Linked Articles