Background Reperfusion of severely ischaemic myocardium can lead to interstitial haemorrhage due to irreversible vascular injury. Haemorrhagic infarction can be detected in vivo by cardiovascular magnetic resonance (CMR) as hypointense signal on T2-weighted (T2W) and T2* imaging. However, the clinical implications of myocardial haemorrhage following acute myocardial infarction (AMI) remain undetermined.

Aims To assess whether the presence of myocardial haemorrhage, as determined by T2-weighted and T2* CMR, influences infarct size, amount of salvaged myocardium and left ventricular ejection fraction (LVEF) following primary percutaneous coronary intervention (PPCI) for AMI.

Methods Forty-eight patients with first presentation acute ST elevation MI, treated successfully with PPCI, underwent CMR imaging within 72 h of admission and at 3 months follow up. The CMR protocol was identical for all scans and included cine imaging, T2W and dual-echo T2* sequences and late gadolinium enhancement (LGE). CMR images were analysed by two experienced observers. The area at risk (AAR) (on T2W) and infarcted tissue (on LGE) were identified using a semi-automated algorithm highlighting myocardium with a signal intensity >2 SD above that of remote normal myocardium. Myocardial haemorrhage and microvascular obstruction (MO), respectively, were identified as regions of hypoenhancement within these areas of hyperenhancement. Myocardial haemorrhage was confirmed by the presence of hypointense signal on T2* imaging.

Results Twenty-six (54%) patients had evidence of MO on LGE. 11 (23%) patients also had myocardial haemorrhage suggested by T2W imaging and confirmed by T2* imaging. Patients with myocardial haemorrhage had greater %LV scar than patients with MO only and patients without MO (overall p<0.001 both at baseline and at 3 months). Myocardial haemorrhage was also associated with smaller %AAR-salvaged myocardium (overall p<0.001), lower %LVEF (overall p=0.001 at baseline and p<0.001 at 3 months) and greater LV end diastolic volume at 3 months (overall p=0.001). Interestingly, patients with myocardial haemorrhage had longer door-to-balloon times than patients with MO only and patients without MO (overall p=0.036). However, there was no overall difference in ischaemia times (i.e. pain-to-balloon) between the groups (overall p=0.74). These results are summarised in Abstract 77 Table 1.

Conclusions Reperfusion haemorrhage following acute myocardial infarction can be detected by CMR and is associated with larger infarct size, diminished myocardial salvage, lower LVEF and adverse ventricular remodelling. The presence of myocardial haemorrhage does not appear to be related to duration of ischaemia. Larger studies in this area are required in the future to investigate the prognostic implications of these findings.