Background

CMR can detect myocardial ischemia by first-pass perfusion and by oxygenation-sensitive CMR (OS-CMR) imaging. While the former can reliably determine myocardial blood flow, the latter integrates other determinants of myocardial oxygenation. Simple breathing maneuvers can trigger a coronary vascular response, which can be monitored by OS-CMR imaging.

Methods

We studied 24 healthy volunteers (37 ± 12 years; 62.5% men) in a clinical 3T MRI system. Each exam included three sets of first pass perfusion images, (1) at rest and, after 1 minute of hyperventilation during (2) a short breath-hold (SBH) or (3) a long breath-hold (LBH), performed in random order. A reader blinded to the maneuver applied, analyzed signal intensity upslope, upslope index and time between 20 and 80% of maximal signal. For inter-observer variability, a different, blinded, reader repeated the analysis in 4 volunteers.

Results

Demographics and LV function data are presented in Table 1. All volunteers tolerated the breathing maneuvers well and completed the study protocol. The average upslope at rest was 1.34 ± 0.58, and increased by 39% during the SBH (1.86 ± 0.70; p < 0.05), diminishing to 1.77 ± 0.82 at the LBH step. The upslope started at 13.8 ± 5.5 and 49.5 ± 7.3 seconds of breath-hold, respectively, on SBH and LBH. Figure 1 shows the relationship between time of breath-hold after hyperventilation and both the individual values of up-slopes and rate-pressure products (RPP). The upslope curve shows two peaks, a early one (15 seconds) coinciding with the peak of the RPP curve; a second one at about 50 seconds, not promoted by the RPP. The upslope index, which accounts for the arterial input, was higher at this late step (rest: 0.077 ± 0.016; SBH: 0.083 ± 0.015; LBH: 0.095 ± 0.019; p < 0.01), as was the myocardial perfusion reserve index (1.25 ± 0.22 vs. 1.09 ± 0.17). In a multiple regression model that included gender, RPP, breath-hold time, caffeine intake, BSA-indexed mass and set order, only gender, RPP and breath-hold time were independently and significantly related to the upslope (R= 0.771; p < 0.001). A different reader repeated the analysis in 4 volunteers; the intra-class correlation for the up-slope was excellent, of 0.990 (95% CI: 0.943-0.997; p < 0.001).

Table 1 Sample demographics and left ventricular function
Full size table
Figure 1
figure 1

Individual values of upslope and rate-pressure products trough time of breath-hold. The upslope curve (filled circles) shows two peaks.

Full size image

Conclusions

The blood flow response to simple breathing maneuvers can be demonstrated by first-pass perfusion CMR, with a early peak dependent on RPP increase, and a late peak due to the vasodilatory effect of long breath holds. Confounding effects of breathing may also have implications for CMR first-pass perfusion imaging performed with pharmacological vasodilators.

Funding

None.