Statistics from Altmetric.com
Editor,—We read with interest the article by Al-Mohammad and colleagues on the prevalence of hibernating myocardium in patients with ischaemic left ventricular dysfunction.1Although we agree with the authors that positron emission tomography (PET) can provide very accurate information for the identification of hibernating myocardium, general statements regarding PET as being the “gold standard” for the diagnosis of viability might be misleading in view of recently published data.2 In a particular subset of patients (those with severe postischaemic left ventricular dysfunction) PET is more accurate than other imaging techniques and, because of its technical characteristics, it is able to provide superior information on tissue viability. However, this requires the adoption of steady-state and standardised study conditions such as those achieved during hyperinsulinaemic euglycaemic clamp for an accurate quantification of the uptake of 18F fluorodeoxyglucose (FDG) by the myocardium. In addition this method does not require the simultaneous measurement of myocardial blood flow for the assessment of viability.
Using this method we have shown that in patients with severe ischaemic heart failure, dobutamine stress echocardiography and PET have similar positive predictive values (68% and 66%) in the identification of hibernating myocardium, but that dobutamine stress echocardiography has a significantly lower negative predictive value than FDG-PET (54%v 96%; p < 0.0001).3 This difference from previously published data reflects the study population in question who had severe left ventricular dysfunction. We have also shown that the baseline ejection fraction influences the predictive accuracy of PET, with the highest positive predictive accuracy in patients with an ejection fraction < 30%.4
We therefore advocate that in patients with postischaemic left ventricular dysfunction, myocardial viability is first sought by use of an easily available and inexpensive technique such as dobutamine stress echocardiography. If this is negative, then a more sensitive method of assessment should be used before ruling out coronary revascularisation. We feel that this should be quantitative PET using FDG during euglycaemic hyperinsulinaemic clamp.5
Regarding the assessment of myocardial blood flow to chronically dysfunctional but viable myocardium (hibernating) in patients with coronary artery disease, we believe that the use of 13N ammonia as a flow tracer may lead to misleading pathophysiological hypotheses. The flow measurements obtained with this tracer reflect the average perfusion in a given mass of myocardium and are not weighted for the presence of scar. In contrast, myocardial blood flow assessed using 15O labelled water mainly reflects flow to well perfused and hence viable tissue as the model used for the calculation assumes that the diffusion of water in scar tissue is negligible. Several papers have shown using PET with 15O labelled water that the resting blood flow to hibernating myocardium is within normal limits. Moreover, similar conclusions have been reached using PET with13N ammonia in patients with hibernating myocardium, but without previous infarction.6
This letter was shown to the authors, who reply as follows:
Barnes and Camici raise several interesting points. First, although dobutamine stress echocardiography and PET have comparable positive predictive accuracy in the assessment of myocardial viability, PET’s superiority was maintained in terms of its negative predictive accuracy.1-1 In addition, our study cohort had severe left ventricular impairment,1-2 a group of patients for whom Barnes and Camici accept that PET is more accurate and provides superior information on tissue viability compared with other imaging techniques. Therefore, in the scientific search for the prevalence of hibernating myocardium in patients with severely impaired left ventricles, PET remains a superior research method. This, however, does not mean that PET should be used routinely in clinical practice. We share with Barnes and Camici the opinion that inexpensive and widely available techniques should be used first, while reserving the more sensitive PET for the negative cases.
Second, with regards to the use of euglycaemic hyperinsulinaemic clamp,1-3 we accept that the ischaemic myocardium is insulin resistant.1-4 Therefore, it might be expected that establishing a euglycaemic hyperinsulinaemic state (glucose clamp) may be associated with better detection of hibernating areas in which glucose uptake is normal or increased. In that respect, Knuutiet al found image quality was superior and fractional utilisation rates of 18F FDG were twice as high during insulin clamp than after glucose loading (p < 0.0001).1-5 However, this technique does not alter18F FDG uptake patterns in different myocardial areas compared to the standard glucose loading protocol.1-5Therefore, there is no evidence to suggest that obtaining superior image quality would have resulted in a different estimation of the prevalence of hibernating myocardium in the cohort of consecutive patients with severely impaired left ventricular contraction.
Third, while we acknowledge the results of perfusion studies using15O labelled water, Barnes and Camici accept that these could not be reproduced using 13N ammonia in patients with previous infarction.1-6 Most of our patients had documented myocardial infarction,1-2 hence it was legitimate to use a widely tested and accepted definition of hibernating myocardium based on the experience gained from 13N ammonia studies. In addition, recent studies confirm the reliability of 13N ammonia as a perfusion marker and suggest its potential use as a predictor of viability.1-7