Impaired left atrial function during exercise cardiac magnetic resonance in patients with heart failure and preserved ejection fraction

Background: A hallmark of heart failure with preserved ejection fraction (HFpEF) is the disproportionate increase in left ventricular (LV) filling pressure during exercise, which may occur in the absence of resting symptoms. Previous study has demonstrated that changes in LA long-axis length (named LA long-axis strain, or LA LS) during exercise, as assessed by real-time Cardiac MRI (RT-CMR) provides prognostic stratification in HFpEF(1). However, LA LS changes reflects changes along a single dimension of a three-dimensional structure. The present study aims to compare the change of LA volumetric function and LA LS during exercise stress using RT- CMR.

Methods: Symptomatic adult patients with a clinical likelihood of HFpEF, determined by H2FPEF(2) and HFA-PEFF scores(3) or asymptomatic individuals with diastolic dysfunction as identified by imaging or invasive right heart catheterization were enrolled. Nineteen healthy adults served as controls. CMR was performed on a 3T scanner (MAGNETOM Vida, Siemens Healthineers, Germany). Exercise was conducted using an MRI compatible supine ergometer (Lode, the Netherlands) starting at 10w and increasing by 10w every 2 minutes until participants reached either >= 85% age-predicted-maximum heart rate or a rate of perceived exertion (RPE) scale of 17-19 (out of 20), or were unable to continue due to dyspnea. RT-CMR cine were acquired at baseline and during peak exercise in 2- and 4-chamber long axis views as well as in a short-axis stack. LA volumes were measured using biplane Simpson’s methods from 2- and 4- chamber views. Difference between baseline and exercise, and between groups were analyzed using independent samples or paired samples T test as appropriate. LA LS was calculated by end-diastolic (ED_length) and end systolic length (ES_length): (ED_length-ES_length)/ED_length x 100(1).

Results: A total of 9 HFpEF patients were enrolled (Tab 1). Both baseline and stress RT-CMR images were successfully acquired in all subjects. Patients with HFpEF achieved significant lower exercise load compared with volunteers (42±26w vs 93±25w, p< 0.001). At baseline, LA LS was markedly lower in HFpEF patients (22.5±9.4% vs 33.3±6.3%, p=0.009); however, LA emptying fraction (EF) was not significantly different (55±19% vs 64±6%, p=0.075). During exercise, HFpEF patients demonstrated blunted augmentation of LA function both on volumetric and long-axis changes (LAEF from 55±19% at baseline to 56±16% at peak and LA LS from 22.5±9.4% to 27.9±11.5%, p all >0.05). In contrast, volunteers exhibited a marked increase in LAEF (from 64±6% to 75±5%, p< 0.001) and LA LS (from 33.3±6.3% to 38.6±5.1%, p< 0.001). The change of LAEF was significantly different between the cohorts (Fig1, Tab1).

Conclusion: While resting LAEF did not differ significantly between HFpEF and volunteers, patients with HFpEF failed to exhibit the expected physiological augmentation of LA function during exercise as shown by the change of LAEF but not by LA LS.

Table 1 Clinical and CMR characteristics of HFpEF and volunteers

	Volunteer (n=19)	HFpEF (n=9)	p value
Age, y	40±15	67±8	< 0.001
Male, n(%)	9 (47.3)	4 (44.4)	0.604
BMI, kg/m2	24.5±4.4	28.0±3.2	0.040
Exercise load (watts)	93±25	42±26	< 0.001
Baseline LAEDVi (ml/m2)	31±7	47±20	0.005
Baseline LAESVi ml/m2	11±3	24±25	0.029
Baseline LAEF(%)	64±6	55±19	0.075
Baseline LA LS (%)	33.3±6.3	22.5±9.4	0.001
Peak LAEDVi (ml/m2)	37±9*	54±17	0.002
Peak LAESVi (ml/m2)	9±3*	26±22	0.002
Peak LAEF (%)	75±5*	56±16	< 0.001
Peak LA LS (%)	38.6±5.1*	27.9±11.5	0.002
Change of LAEDVi (ml/m2)	6.3±6.6	7.5±12.6	0.741
Change of LAESVi (ml/m2)	-1.7±2.7	1.9±7.9	0.083
Change of LAEF (%)	10.5±6.0	0.2±8.4	< 0.001
Change of LA LS (%)	5.3±5.5	5.5±11.1	0.954

Figure 1 LA size and function changes during exercise; *: p<0.05.