In this study of a longitudinally-followed community-dwelling cohort initially free of clinical CVD, we found that decreased MCF, a novel volumetric LV index, predicts hard CVD events over > 5-year follow up. The lowest quartile of MCF was predictive of a major adverse hard CVD event after adjustment for multiple CVD risk factors, including age, systolic blood pressure, dyslipidemia, diabetes, treatment for hypertension or hyperlipidemia, and smoking, as summarized by the FCRS. (The FCRS does not incorporate any CMR results.) Further, decreased MCF remained predictive of future hard CVD events after additional adjustment for LV mass. We also found greater MCF in women versus men.
LVEF is widely used to identify abnormal ventricular function, to perform risk stratification and to guide therapy. However, normal-range LVEF does not necessarily imply low CVD risk, or even freedom from symptomatic CVD. Indeed, heart failure with normal-range LVEF is a common clinical scenario which carries an [6
]. One means by which LVEF may be preserved despite myocardial contractile dysfunction is via concentric remodeling, where wall thickness increases relative to cavity size [9
]. Although these concentric changes in LV geometry may allow normalization of the LVEF, concentric remodeling or LV hypertrophy are both associated with excess hard CVD events, as compared to normal LV geometry [11
]. Given that MCF is the inverse of LV mass multiplied by LV stroke volume and a fixed scaling factor, we accordingly investigated the impact of adjusting for LV mass and found that MCF remained predictive of future hard CVD events. Further, depressed MCF may identify an additional subset of at-risk individuals as compared with increased LV mass alone. In this study cohort, 10% of the hard CVD participants had low MCF but also low LV mass.
Considering MCF mathematically, it can also be expressed as 1.05 × [LVEDV/LVM – LVESV/LVM], where ESV is end-systolic volume. This indicates that MCF is, apart from an overall scaling factor of 1.05 g/ml, the inverse of LV concentricity (LVM/LVEDV ratio) from which the ratio of ESV to LV mass is subtracted. Thus an increased concentricity means a decreased MCF, but MCF will decrease further with larger ESV. This interpretation of MCF suggests that it might be considered an index which encompasses both LV concentricity and systolic function (worse function leading to larger ESV and thus a greater decrease in MCF than would be based on LV concentricity alone). That concentricity is an important component of MCF appears to be reflected by our finding that lowest-quartile MCF was no longer an independent predictor of hard CVD events in a model containing both MCF and concentricity. However, MCF nonetheless trended toward significance, even after adjustment for concentricity, and it may be that lack of statistical power limited our ability to find independent predictive value for MCF in this model.
MCF does not require any additional CMR image acquisition or analyses above those already performed to determine LV mass and volumes and importantly, scanner workflow is not altered. Data acquisition for LV function typically required < 10 minutes in this study using a single breath-hold per image slice, and faster imaging methods are now available [14
]. Manual delineation of contours requires approximately 6 minutes, and semi-automated methods can reduce this time further [16
]. Computationally, MCF is straightforward, as it is the ratio of stroke volume to myocardial volume. Myocardial volume is related to myocardial mass by a simple scaling factor, the specific gravity of myocardium (1.05 g/cc) but is computationally advantageous over LV mass in that it allows MCF to be a dimensionless ratio, as is LVEF.
The Framingham Heart Study population is largely Caucasian and the Offspring cohort is middle-aged or older; our results may not extend to other ethnic or age groups. Our study used an older (though current at time of CMR scanning) segmented K-space gradient-echo cine method, while LV functional imaging is now performed using steady-state free-precession (SSFP) cine sequences. Compared with gradient-echo, SSFP yields greater LV volumes and lower LV mass [17
], and thus higher MCF values. As SSFP is now the overwhelming sequence of choice for assessment of LV size and function, we do not report cut-points for MCF based on our gradient echo data as these would be of little clinical or research value.
In this study, decreased MCF was an independent predictor of future hard CVD events, but the relatively modest number of events during follow-up was insufficiently powered for detailed analyses of the subtypes of hard CVD events. On sex-specific analyses, the difference in MCF between hCVD+ and hCVD- participants was significant only in women, most likely due to sample size and number of events rather than sex-specific predictive value of MCF. We found significantly lower MCF in men than women, but whether sex-specific criteria for identification of depressed MCF are needed is unknown and should be examined in a larger cohort. If CMR is used, this should be done using SSFP imaging. Importantly, MCF is not CMR-specific. Indeed, it was initially described in the context of three-dimensional (3D) echocardiography [1
]. King et al. suggested that MCF should be determined using a volumetric (3D) technique, due to the greater accuracy and reproducibility of 3D, vs. 2D, methods for determination of ventricular and myocardial volumes. However, it remains to be investigated as to whether MCF by 2D methods predicts future hard CVD events, as we did not determine biplane CMR LV volumes and mass in the present study. Possible additional relationships between LV mass, concentric remodeling and MCF await further investigation, particularly with regard to persons with both low MCF and low LV mass, but the number of such participants in the present study was too small to permit us to explore this in detail. Also, whether MCF provides additional prognostic value when echocardiographic diastolic function parameters are available also remains to be determined, as there were no contemporaneous echocardiographic data for this study. Finally, the value of MCF in patients who already have clinically overt CVD or depressed LVEF, and whether changes in MCF have prognostic value remains to be investigated.
MCF is a novel, easily-determined volumetric LV index. Decreased MCF is an independent predictor of future CVD morbidity, even in the presence of normal LVEF and after adjustment for both traditional CVD risk factors and LV mass.