Abstract
Background
Myocardial deformation measured by myocardial strain is an important marker of cardiovascular disease (CVD). We aimed to establish normal values and examine age- and sex-differences in left and right ventricular (LV, RV) strain using feature tracking cardiovascular resonance (FT-CMR) in adults free of CVD in the community.
Methods
Framingham Heart Study Offspring participants without CVD, hypertension, or diabetes completed CMR (1.5T) (n = 903, 59% (533/903) women, 36–88 years of age). Global longitudinal, circumferential, and radial strain (GLS, GCS, and GRS, respectively) were measured by semi-automated LV and RV myocardial segmentation of short-axis, 2-chamber, and 4-chamber images with cvi42 (v.5.13, Circle Cardiovascular Imaging, Calgary, Canada).
Results
Mean LV GLS, GCS, and GRS were −17.8 ± 2.5%, −19.7 ± 2.4%, and 34.9 ± 6.9%, respectively; RV GLS, GCS, and GRS were −20.4 ± 4.0%, −11.7 ± 3.5%, and 19.1 ± 6.5%, respectively. All strain values were of greater magnitude in women than men (all p<0.005). LV and RV GCS and GRS were of greater magnitude (more negative and more positive, respectively) with increasing age (all p<0.001), but GLS was not associated with age. Stratified analysis by sex showed LV and RV GCS and GRS were greater in magnitude with increasing age in women (all p<0.001), whereas no age associations in men were observed.
Conclusion
We provide FT-CMR reference values for LV and RV strain in a large, community-dwelling U.S. adult cohort free of CVD and CVD risk factors. Women demonstrated greater magnitude LV and RV GCS and GRS with increasing age. The clinical implications of sex-differences in ventricular strain and mechanical remodeling with age deserve further study.
Key Words: Ventricular strain, Cardiovascular magnetic resonance, Feature tracking
Graphical abstract
1. Introduction
Measuring myocardial deformation, myocardial strain is an important subclinical marker of myocardial dysfunction in cardiomyopathy and predicts all-cause mortality in patients with heart failure and valvular heart disease [1], [2], [3], [4]. Myocardial strain also predicts mortality and adverse cardiovascular outcomes in patients with congenital heart disease and after myocardial infarction [5], [6]. Moreover, left ventricular (LV) global longitudinal strain is used in the detection of early cancer treatment-related myocardial toxicity and is incorporated into current guidelines for chemotherapy monitoring [7]. Overall, LV strain is a more sensitive marker of early and subclinical myocardial dysfunction compared with other measures of cardiac function such as ventricular ejection fraction [8]. Cardiovascular magnetic resonance (CMR) is the gold standard non-invasive method to assess cardiac structure and function. In addition to comprehensive characterization, feature tracking CMR (FT-CMR) may be easily applied to routine cine images to assess myocardial strain [9]. While FT-CMR assessment of ventricular strain in prior years suffered from variability between readers, greater consistency in modern analysis packages allows more reliance on these tools for both clinical and research applications [10]. Thus, establishing reliable reference values is paramount to result interpretation.
Prior work reporting normative values of LV and right ventricular (RV) myocardial strain have been limited by cohort and sample size [11], [12]. Additionally, few studies have examined patterns of FT-strain in the general population. We aimed to address this knowledge gap by analyzing age- and sex-differences in FT-CMR LV and RV strain in a free-living community cohort without CVD.
2. Methods
2.1. Study population and clinical covariates
The Framingham Heart Study (FHS) Offspring cohort, including the children (and their spouses) of the original FHS cohort, is a community-dwelling sample of adult men and women who have received detailed history, physical examinations, and lab work every 4–6 years beginning with Exam 1 in 1971. The sampling design, characteristics, and examinations of the FHS Offspring study have been previously described [13]. Offspring cohort members participating in Exam 7 and who lived in Massachusetts or a contiguous state were invited for CMR using a random sampling strategy to recruit from strata of gender, age decile, and quintile of Framingham Risk Score [14]. Potential study participants were excluded for contraindications to CMR (e.g., pacemaker, implantable cardioverter-defibrillator, history of metallic devices or foreign bodies, severe claustrophobia) or for non-sinus rhythm. Of 3539 participants who attended Exam 7, 1858 participants completed CMR testing with strain analysis of ventricular myocardium. To study a healthy sample without CVD risk factors for the present study, we excluded those with hypertension (HTN), diabetes mellitus (DM), and prevalent CVD (coronary artery disease, heart failure, stroke, transient ischemic attack, peripheral artery disease), resulting in a final sample of 903 participants. The final cohort consisted entirely of White participants. This group was subdivided into 4 age groups of similar sample size (≤50 years, 51–60 years, 61–70 years, and ≥71 years) for comparison. Clinical covariates from Exam 7, immediately proximate or contemporaneous to CMR, were used in the present study. The measurement and definition of covariates used in the present study, including body mass index (BMI), HTN, and DM, have been described previously [15].
The study protocol was approved by the institutional review boards of both Boston University Medical Center and Beth Israel Deaconess Medical Center. All study participants provided written informed consent.
2.2. CMR imaging
Non-contrast CMR was performed on a 1.5T CMR scanner (Gyroscan ACS NT, Philips Medical Systems, Best, The Netherlands) with a 5-element commercial cardiac array receiver coil as previously described [15]. Electrocardiogram-gated, balanced steady-state free precession (bSSFP) images were acquired during end-expiratory breath holds in the LV and RV short-axis (SAx) orientation spanning the entire ventricle as well as in the LV 2-chamber (2Ch) and 4-chamber (4Ch) planes (Fig. 1) (repetition time [TR] = 3.2 ms, echo time [TE] 1.6 ms, flip angle 60°, field-of-view (FOV) 400 mm, matrix 208 × 256, slice thickness 10 mm, gap = 0).
Fig. 1.
Endocardial (red) and epicardial (green) border segmentation for FT-strain analysis. LV and RV endocardial and epicardial contours in 2Ch (top left), 4Ch (bottom left), and SAx (right) views. FT feature tracking, LV left ventricle, RV right ventricle, 2Ch two chamber, 4Ch four chamber, SAx short axis
2.3. CMR analysis
CMR data analysis was performed using commercial software (cvi42 v.5.13, Circle Cardiovascular Imaging, Calgary, Alberta, Canada) by an operator (KP, 3 years CMR analytic experience) blinded to all clinical data. From bSSFP short-axis cine images, a semi-automated method was used to trace the endocardial and epicardial LV borders at end-diastole (first images of the cine data set) and end-systole, as described previously (Fig. 1) [16]. Basal slices containing ≥50% of myocardial circumference were considered for analysis, with manual adjustment of the endo- and epicardial contours to exclude regions external to the LV (particularly the LV outflow tract and left atrium), using the 4-chamber and 5-chamber cine SSFP images at end-diastole and end-systole as reference. LV mass (LVM) was calculated by software as the area between end-diastolic and end-systolic contours multiplied by the slice thickness and the density of myocardial tissue (1.04 g/mL), and summed for all myocardial slices. A similar summation of disks method was used to calculate LV cavity volumes at end-diastole and end-systole, with calculation of LV ejection fraction (LVEF). LVM, LV end-diastolic volume, and LV end-systolic volume were divided by body surface area (BSA) to index to body size (LVMI, LVEDVI, and LVESVI, respectively). A similar method for the RV was employed to compute RVMI, RVEDVI, RVESVI, and RVEF, respectively. All LV and RV short-axis endocardial contours were manually adjusted as necessary to exclude papillary muscles and trabeculations, which were both considered part of LV and RV cavity volume.
Additional endocardial and epicardial contours of long axis (2Ch and 4Ch) cine bSSFP images of the LV at end-diastole were traced in similar fashion to short-axis images for computation of LV myocardial strain using the FT module of cvi42. RV FT analysis was performed using the endocardial border of the RV free wall in SAx and 4Ch orientations. Myocardial endocardial and epicardial deformation was tracked throughout the cardiac cycle (end-systolic frame shown in Fig. 2). FT-CMR Global strain values are reported as the peak values of the mean strain curves of the LV and RV in three established orientations (global longitudinal, circumferential, and radial strain [GLS, GCS, GRS, respectively]) A set of color maps of GLS, GCS, and GRS in the 17-segment model and the corresponding strain graphs are shown in Fig. 3.
Fig. 2.
Deformation of myocardial endocardial (red) and epicardial (green) border tags at end-systole. Movement of tags and tracking of myocardial strain in 2Ch, 4Ch, and SAx images are shown. 2Ch two chamber, 4Ch four chamber, SAx short axis
Fig. 3.
FT-CMR color maps and graphs of GLS, GCS, and GRS (left-right). Top panel: 17-segment model color overlay map of GLS, GCS, GRS, respectively shown left to right. Bottom panel: corresponding GLS, GCS, and GRS strain curves. GLS global longitudinal strain, GCS global circumferential strain, GRS global radial strain
2.4. Reproducibility
Inter-observer variability was tested in 20 randomly selected participants with assessment of CMR images and strain measurements performed independently by two investigators (K.P., V.J., each with 3 years of dedicated CMR analytic experience).
2.5. Statistical analyses
All continuous variables were normally distributed by quantile-quantile plot and were described as mean ± standard deviation (SD). Means of baseline clinical characteristics and strain values were compared between men and women by independent samples t-test. Mean strain values were compared across four age groups by one-way analysis of variance (ANOVA) with post hoc pairwise comparisons by Tukey’s honest significant differences. We used a two-tailed Bonferroni corrected p value of 0.05/6 = 0.0083 as the threshold for statistical significance to correct for multiple testing (six strain outcome variables). Multivariable linear regression analysis and stratified analysis were performed to determine interaction effects between age and sex on strains. We also used linear regression analysis to examine associations between strain and BSA/body mass index (BMI), and we conducted a secondary analysis of current guideline normotensive and stage I hypertensive subgroups. Inter-observer reproducibility was assessed by calculating mean bias and 95% limits of agreement from Bland-Altman analyses and intraclass correlation coefficients (ICC). Statistical analyses were performed using R (version 4.1.0).
3. Results
3.1. Baseline characteristics of participants
Table 1 presents baseline characteristics of the 903 subjects (59% (533/903) women; 62 ± 9 years; range 36–88 years). There was no significant difference in age between men and women (p = 0.17). Average BMI, BSA, systolic blood pressure (SBP), heart rate (HR), LVMI, RVMI, LVEDV, LVESV, RVEDV, and RVESV were greater in men (all p≤0.002). However, LVEF and RVEF were greater in women (p<0.001).
Table 1.
Clinical characteristics of study sample.
| Characteristics | All (n = 903) | Men (n = 370) | Women (n = 533) |
|---|---|---|---|
| Age, yrs | 62 (9) | 61 (8) | 62 (9) |
| BMI, kg/m2 | 26.6 (4.3) | 27.5 (3.6) | 25.9 (4.5) |
| BSA, m2 | 1.84 (0.21) | 2.01 (0.15) | 1.72 (0.16) |
| SBP, mmHg | 117 (11) | 119 (10) | 115 (12) |
| HR, beats/min | 64 (10) | 63 (11) | 65 (10) |
| Current smoking | 77 (8.5%) | 26 (7.0%) | 51 (9.6%) |
| LV Measures | |||
| LVM index, g/m2 | 46 (9) | 52 (8) | 42 (6) |
| LVEDV index, mL/m2 | 75 (13) | 81 (15) | 71 (11) |
| LVESV index, mL/m2 | 30 (8) | 33 (8) | 28 (6) |
| LVEF, % | 60 (5) | 59 (6) | 61 (5) |
| LVEDD index, mm/m2 | 28 (3) | 26 (3) | 28 (3) |
| IVST, mm | 8 (1) | 8 (1) | 7 (1) |
| RV Measures | |||
| RVM index, g/m2 | 18 (4) | 19 (4) | 17 (3) |
| RVEDV index, mL/m2 | 69 (1514) | 78 (15) | 64 (11) |
| RVESV index, mL/m2 | 32 (10) | 37 (10) | 28 (8) |
| RVEF, % | 55 (9) | 52 (9) | 56 (9) |
All p≤0.002 for men vs. women except age, p = 0.17
Data are presented as mean (SD), except Current smoking presented as count (prevalence). BMI body mass index, BSA body surface area, SBP systolic blood pressure, HR heart rate, LVM left ventricular mass, LV left ventricular, EDV end-diastolic volume, ESV end-systolic volume, EF ejection fraction, EDD end-diastolic dimension, IVST interventricular septal thickness, RVM right ventricular mass, RV right ventricular
3.2. Normal values and age- and sex-differences
Mean LV and RV strain values across four age groups (≤50, 51–60, 61–70, and ≥71 years) are presented in Fig. 4 and Supplemental Table 1. With increasing age, we observed more negative LV GCS and more positive LV GRS (both p<0.001), indicating greater strain magnitude and increased contractility. Pairwise comparisons identified differences between the youngest group (<50 years) and all three older groups for both LV GCS and GRS. No significant age trend was seen for LV GLS (p = 0.23). Similarly, across increasing age groups, we observed more negative RV GCS and more positive GRS (both p<0.001) without significant age trend in RV GLS (p = 0.03). Pairwise testing for RV GCS and GRS showed differences between the youngest group and the two oldest groups as well as between the 51–60 year age group and the two oldest groups (Fig. 4).
Fig. 4.
Left and right ventricular strain across age groups. Boxplots of global strain values across age groups. Brackets indicate significant pairwise comparisons (p<0.0083; †: p<0.005; *: p<0.0001). LV left ventricle, RV right ventricle, GLS global longitudinal strain, GCS global circumferential strain, GRS global radial strain
LV and RV strain values by sex are presented in Fig. 5 and Supplemental Table 2. Compared to men, women had greater magnitude of both LV and RV GLS, GCS, and GRS (more negative GLS and GCS, more positive GRS, all p<0.005).
Fig. 5.
Left and right ventricular strain by sex. Boxplots of global strain values for men and women. All comparisons between sexes were statistically significant (p<0.0083). LV, left ventricle, RV right ventricle, GLS global longitudinal strain, GCS global circumferential strain, GRS global radial strain
3.3. Joint relations of age and sex with ventricular strain
Mean values of LV and RV strain across age groups, among men and women separately, are presented in Table 2 and Fig. 6. Among men, there was no significant trend of LV or RV strain across age groups, and additionally, no pairwise differences between age groups in all strain parameters. There was no significant trend of LV GLS with age in women. In contrast, women had greater magnitude of LV GCS and GRS (both p<0.001) with age, with significant pairwise testing between the youngest and the two oldest groups. Similar to LV strain, we observed a greater magnitude of RV GCS and GRS in higher age groups (both p<0.001) and with pairwise differences between the younger and older groups among women but not significantly in men. The relations of RV GLS with age were less clearly unidirectional. While RV GLS slightly increased in magnitude across age to 70 years, pairwise testing revealed its significance was due to a sharp decline in strain in the >71 year group rather than a consistent trend with increasing age groups.
Table 2.
Left and right ventricular strain across age groups and by sex.
| Men (n = 370) |
Women (n = 533) |
|||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Strain | ≤50 y (n = 43) | 51–60 y (n = 143) | 61–70 y (n = 132) | ≥71 y (n = 52) | p | ≤50 y (n = 37) | 51–60 y (n = 234) | 61–70 y (n = 169) | ≥71 y (n = 93) | p |
| LV GLS, % | −16.5 (2.8) | −16.4 (2.0) | −16.7 (2.2) | −16.8 (2.3) | 0.61 | −18.5 (2.2) | −18.7 (2.3) | −18.7 (2.4) | −18.9 (2.5) | 0.84 |
| LV GCS, % | −17.9 (2.2) | −18.5 (1.9) | −18.6 (2.6) | −18.9 (2.3) | 0.15 | −19.2 (1.3) | −20.3 (2.0) | −20.8 (2.2) | −20.9 (2.1) | <0.001 |
| LV GRS, % | 29.8 (5.9) | 31.2 (5.1) | 32.0 (6.9) | 32.8 (6.3) | 0.07 | 33.3 (3.6) | 36.6 (6.0) | 38.2 (7.0) | 38.7 (6.7) | <0.001 |
| RV GLS, % | −19.4 (4.3) | −19.8 (3.7) | −19.6 (4.0) | −20.2 (4.3) | 0.75 | −19.6 (4.3) | −21.0 (4.0) | −21.5 (3.6) | −19.6 (4.3) | <0.001 |
| RV GCS, % | −10.4 (3.3) | −11.0 (3.3) | −11.6 (3.5) | −11.8 (2.7) | 0.07 | −9.6 (3.0) | −11.3 (3.5) | −12.7 (3.5) | −13.0 (3.2) | <0.001 |
| RV GRS, % | 16.6 (5.9) | 17.5 (6.1) | 19.0 (6.6) | 19.2 (5.4) | 0.04 | 15.7 (4.9) | 18.6 (6.2) | 21.0 (7.1) | 21.6 (6.1) | <0.001 |
Data are presented as mean (SD). All age groups are in years. P values represent one-way ANOVA across age groups. ANOVA analysis of variance, LV left ventricle, RV right ventricle, GLS global longitudinal strain, GCS global circumferential strain, GRS global radial strain, SD standard devi
Fig. 6.
Left and right ventricular strain across age groups and by sex. Boxplots of global strain values across age groups for men (blue) and women (red). Brackets indicate significant pairwise comparisons (p<0.0083; †: p<0.005; *: p<0.0001). LV left ventricle, RV right ventricle, GLS global longitudinal strain, GCS global circumferential strain, GRS global radial strain
3.4. Impact of body surface area and body mass index on ventricular strain
Multivariable linear regression analyses between either continuous BSA or BMI and all strain measures revealed no significant relations (p≥0.01) with the exception of LV GLS (p = 0.002). However, in interaction analysis, there was no effect modification between either BSA or BMI on the relations between sex and any strain measures (all p≥0.3).
3.5. Subgroup analysis using current blood pressure definitions
While the definition of HTN in FHS has for decades utilized SBP ≥140 mmHg or diastolic blood pressure (DBP) ≥90 mmHg, updated hypertension guidelines specify stage I HTN as SBP 130–139 mmHg or DBP 80–89 mmHg [17]. Thus, we performed a subgroup analysis of individuals who are normotensive by updated blood pressure criteria for HTN. After exclusion of those with SBP ≥130 mmHg or DBP ≥80 mmHg, mean LV and RV strain values across age groups separated by sex for the current guideline normotensive subgroup (n = 632) are presented in Supplemental Table 3. With the exception of RV GLS in women, all significant associations observed in the initial analysis remained significant using current lower thresholds for HTN. An analysis of the excluded patients, or the stage I hypertension subgroup (n = 238), are presented in Supplemental Table 4 and similarly showed significantly greater LV and RV strains in women with the exception of RV GCS (p = 0.014). However, the reduced sample size of this subgroup precluded stratification across multiple age groups.
3.6. Inter-observer variability
Inter-observer variability for FT-CMR LV and RV strain is shown in Table 3. All LV and RV strain parameters had an inter-observer ICC >0.9 except for RV GLS (0.70).
Table 3.
Inter-observer variability.
| Strain | Mean (SD) | Limits of agreement | ICC (95% CI) |
|---|---|---|---|
| LV GLS | 0.21 (1.49) | −2.71 to 3.13 | 0.92 (0.80, 0.97) |
| LV GCS | −0.03 (0.70) | −1.40 to 1.34 | 0.99 (0.96, 0.99) |
| LV GRS | 0.09 (2.12) | −4.07 to 4.24 | 0.98 (0.96, 0.99) |
| RV GLS | 0.20 (4.10) | −7.84 to 8.23 | 0.70 (0.38, 0.87) |
| RV GCS | −1.18 (1.48) | −4.08 to 1.72 | 0.92 (0.64, 0.97) |
| RV GRS | 1.89 (2.84) | −3.68 to 7.46 | 0.93 (0.75, 0.97) |
Data are presented as mean (SD), lower to upper limit, or coefficient (95% CI). LV left ventricle, RV right ventricle, GLS global longitudinal strain, GCS global circumferential strain, GRS global radial strain, ICC intraclass correlation coefficient, SD standard deviation
4. Discussion
In this study of an adult cohort free of CVD, HTN, and DM, we provide normative values for global LV and RV GLS, GCS, and GRS by age and sex groups. We demonstrate a trend toward greater magnitude strain with increasing age for LV and RV GCS and GRS that appears largely driven by these observations in women. We confirm excellent reproducibility for all LV and RV strain parameters, with the exception of RV GLS which was moderate. Our study represents one of the largest using FT-CMR for analysis of biventricular strain in healthy, community-dwelling adults free of CVD risk factors and with a broad age range.
4.1. Normative values
Establishing FT-CMR reference values for LV and RV strain has previously met challenges. Limitations due to the study sample have been numerous, including limited sample sizes, limited age groups, hospital-based or referral samples, or inclusion of individuals with CVD or CVD risk factors. Additionally, creation of FT-CMR reference values has been hampered by variability in values obtained due to differences in image technique, software vendors, and computation using peak or mean global strain values [11]. A recent meta-analysis of 3359 healthy subjects from 44 FT-CMR studies identified FT vendor as a significant contributor to the heterogeneity of both LV and RV strain across studies [18]. However, the largest sample in this analysis had only 180 subjects, highlighting relatively small sample sizes remaining a limitation of the literature [18]. Consistent with the literature, our LV strain values were generally lower than those of studies that used TomTec software, supporting the findings of Yang et al. regarding the effects of FT vendors on the generalizability of strain measurements [18], [19], [20], [21]. Recent larger studies have focused on establishing normal values in specific population samples, including a study of 566 healthy Chinese subjects by Li et al.[22], [23], [24], [25]. While our results show good agreement with LV strain values reported by Li et al., Zhang et al., and Valdés et al., all of whom used cvi42 software [22], [24], [26], others have found greater values for LV GLS, GCS, and GRS [27], [28].
As compared with CMR assessment of the LV, RV strain determination has been even more limited [12]. Measurement of the RV is made difficult by its complex anatomy, thin free wall, significant trabeculations, and respiratory variability in filling [29]. Few studies have been published using FT-CMR to measure RV strain values, with these studies also suffering from modest sample sizes [8], [12], [27], [29], [30]. The RV GLS values in our study were lower than those of others using the same software, which could be related to the regional distribution of our study population [8], [12], [22], [27], [29], [30]. Although the meta-analysis by Yang et al. did not find regional distribution effects on RV GLS, they did not include any North American studies [18]. When compared against Asian cohorts, our RV GCS and GRS values were in line with those reported by Hwang et al. but lower than those reported by Li et al. and Liu et al. [22], [27], [29]. The observed variability in RV strain even among studies using cvi42 software suggests effects from outside variables, such as regional population differences. Our study provides one of the largest samples free of CVD and CVD risk factors, to establish normal values for LV and RV strain.
It is important to recognize that absolute strain values may differ between FT-CMR and other methods of CMR strain assessment such as tagged CMR, displacement encoding with stimulated echoes, phase contrast CMR, and fast strain-encoded (fSENC) CMR [9]. While our study did not compare techniques, literature comparing values obtained by different methods of strain among the same patients has demonstrated variations in absolute values. While LV GCS may be comparable between FT-CMR and both fSENC and tagged CMR, considered the reference standard for GCS, these techniques may not yield similar values for LV GLS [9], [26]. Finally, when compared with speckle tracking echocardiography, FT-CMR has been shown to correlate well but is not interchangeable for LV and RV strains in both healthy subjects and those with CVD [31], [32], [33].
4.2. Age- and sex-relations with myocardial strain
Our findings suggest that greater age is an important determinant of circumferential and radial strain, but not GLS, similar to other studies [19], [21], [22], [23], [26], [27]. Our observations of greater LV GCS and GRS with advanced ages are consistent with results from other studies that utilized cvi42 software [22], [24], [26], [27], [28]. Notably, variable results of the relations between age and strain were reported in studies using different software for strain analysis [19], [21], [23]. In parallel to LV GCS and GRS, we observed that RV GCS and GRS increased with age, which corroborates findings by Liu et al. and Li et al [22], [29]. The maintenance of RV GLS through age, observed also by others [8], [12], [23], [29], [30], may be notable in the preservation of RV systolic function throughout life, given the major contribution of longitudinal shortening with RV contractility. Meanwhile, age-related changes in RV GCS may be of clinical interest in conditions with RV pressure overload, such as with pulmonary hypertension, congenital heart disease, and post-cardiac surgery, where RV contractility shifts from predominantly longitudinal to circumferential shortening [34], [35], [36].
We identified differences in strain between sexes, with women having greater GLS, GCS, and GRS for both ventricles. Among the subset with stage I hypertension, we observed similar strain values to the overall cohort, including greater LV and RV strains in women. There were no significant relations between body size as estimated by BSA or BMI and all strain measures except LV GLS and BSA. However, interaction analyses between either BSA or BMI on the relations between sex and all LV and RV strain measures suggested that sex-differences in strain were not due to differences in body size. While a weak association has been previously reported between BSA and LV GLS, GCS, and GRS [37], the lack of significant effect modification by body size in the relations between sex and strain is consistent with our understanding of other measures of systolic function, including LVEF.
Although meta-analyses by Vo et al. and Yang et al. found no association of LV GLS, GCS, and GRS with sex, many individual studies have shown sex-differences in LV strain consistent with our results [11], [18]. Notably, Li et al., Valdes et al., and Hwang et al. also reported greater LV GLS, GCS, and GRS in women [22], [26], [27]. Additional studies have also found greater LV GLS in women but reported more variable LV GCS and GRS findings, with some showing only a trend towards greater LV GCS in women and others showing significantly greater LV GRS in men [19], [20], [21], [23], [24]. In general, LV GRS is less accurately measured by FT algorithms and often has larger standard deviations which may contribute to the discrepancies in sex associations across studies. The sex-differences in RV strain we observed were concordant with Li et al., Hwang et al., and Liu et al., who also showed that women had greater RV GLS, GCS, and GRS than men, although RV GLS did not reach significance in the latter study [22], [27], [29]. Whereas, some others identified no sex-differences in RV GLS [8], [23], [30]. Overall, our findings of greater LV and RV strain in women support the need for sex-specific reference values. Our stratified analyses suggest that the observed increases in LV and RV GCS and GRS with age were largely driven by differences in women. While strain measures were of greater magnitude with advancing age in men, the lack of statistical significance may be related to both smaller sample sizes as well as less difference in mean strain values per age group. Thus, while men may also have greater magnitude of LV and RV GCS and GRS with greater age, the magnitude of this increase seems lower than in women. This may help to explain why multivariable linear regression interaction analysis between age and sex on strain was not significant in our study.
4.3. Clinical implications
In our study, women demonstrated greater increase in the magnitude of LV and RV GCS and GRS with older age, implying sex-differences in age-related ventricular mechanical remodeling. A majority of patients with heart failure who are elderly, especially women, have preserved ejection fraction, yet understanding of the underlying pathophysiology remains limited [38]. Even among normal subjects and independent of their smaller chamber size, older women tend to have higher LVEF [38]. A CMR study of the multi-ethnic study of atherosclerosis (MESA) cohort showed many changes in LV structure and function with age including an increased mass to volume ratio, enhanced ejection fraction, and abnormal diastolic strain [39]. Furthermore, a three-dimensional echocardiography study of the Atherosclerosis Risk in Communities (ARIC) cohort demonstrated lower myocardial diastolic relaxation, reduced LV GLS, and borderline greater LV GCS with age [40]. Our data support the theory that age-related LV stiffness and diastolic dysfunction are compensated for by an increase in systolic wall thickening of circumferential and radial fibers, and this pattern of compensatory remodeling may be especially prominent in women. Among hypertensive patients, it has been found that the predominant pattern of hypertrophy in women was concentric as opposed to eccentric in men [38]. Future studies are needed to investigate these sex-differences in ventricular remodeling with age and associations with the development of heart failure in older women, and separately address the influence of contemporary medications that modulate heart failure risk and cardiac function.
4.4. Reproducibility of FT-CMR strain
Our mean inter-observer variability was overall low though wider limits of agreement were present for RV strains and LV GRS. Our reproducibility by ICC was excellent for all LV and RV strains with the exception of RV GLS, which was moderate. Similarly, Li et al. and Hwang et al. found lower reproducibility for RV strain, in particular RV GLS, underscoring the difficulty of RV strain measurement due to its anatomy [22], [27]. Additionally, comparatively low reproducibility of radial strain has been shown in prior studies and may be related to the need to simultaneously track endocardial and epicardial borders to derive radial strain [21]. Overall, our study found strong relative consistency and reproducibility, but the greater differences in absolute RV strain and radial strain values could suggest some systematic inter-observer bias.
5. Limitations
Though our analyses are conducted in a large community sample, several limitations bear consideration. We sought to determine normative values and the influence of key determinants of age and sex on ventricular mechanics in older adults. Two intrinsic corollaries result, including (a) CVD- and risk factor-free survival bias in women, resulting in a larger sample size of older women in our study, and (b) the possible presence of other subclinical CVD risk factors that could influence strain. Despite elimination of major risk factors including hypertension, diabetes, and prevalent CVD, unmeasured confounders may remain. Although our initial cohort included participants with what is presently considered stage I hypertension, our normotensive subgroup analysis demonstrated similar results in the significant relationships originally identified. Moreover, generalizability of our results to younger patients may be limited by the relatively small number of participants less than 45 years old. It is worth noting that the sex stratified age subgroups, in particular the youngest and oldest groups, had relatively fewer participants, which can limit statistical power and the ability to detect significant pairwise differences in our study. Regarding considerations of our analysis method, we utilized cvi42, a predominant clinical CMR software analysis tool worldwide. Inherent to FT-CMR are limitations related to image quality and accurate tracking of endocardial borders. Although reasonably high in our study, temporal resolution has also been shown to influence the accuracy of strain calculation [26]. Finally, the predominance of White adults of Western European descent in our sample may restrict generalizability, and further study into the influence of race and ethnicity on strain measures are needed.
6. Conclusion
In summary, in a large community-dwelling adult cohort free of CVD risk factors who underwent CMR, we provide normative age- and sex-specific FT-CMR reference values for LV and RV strain measures. While our results demonstrated greater magnitude ventricular strain with older age, stratified analysis revealed these trends were sex-specific, in women but not men. Further investigation is needed to understand the clinical implications and underlying myocardial mechanics of these group differences in ventricular strain.
Funding
The Framingham Heart Study is supported by the National Heart, Lung and Blood Institute (NO1‐HC‐25195 and HHSN268201500001I). This work was also supported by NHLBI R01‐HL70279 and R01HL155717.
Author contributions
Rohit Pillai: Writing – review & editing, Writing – original draft, Visualization. Lu Zhang: Visualization, Formal analysis, Data curation. Kevin Peters: Validation, Formal analysis, Data curation. Vibhav Jha: Validation, Data curation. Christopher J. O’Donnell: Writing – review & editing, Conceptualization. Warren J. Manning: Writing – review & editing, Supervision, Conceptualization. Connie W. Tsao: Writing – review & editing, Supervision, Project administration, Methodology, Funding acquisition, Conceptualization.
Disclosures
The authors report no financial disclosures related to the content of this research. CJO is an employee of Novartis Biomedical Research, Cambridge, Massachusetts, USA.
Declaration of competing interests
The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Christopher J. O’Donnell reports a relationship with Novartis Institutes for BioMedical Research Inc. that includes employment. The other authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgements
The authors wish to thank the staff and participants of the Framingham Heart Study. Data supporting this study are available on request from dbGaP Authorized Access at phs003593.v1.p1. Data are not publicly available to protect the privacy of research participants.
Footnotes
Supplementary data associated with this article can be found in the online version at doi:10.1016/j.jocmr.2025.101902.
Appendix A. Supplementary material
Supplementary material
.
References
- 1.Maceira A.M., Prasad S.K., Khan M., Pennell D.J. Reference right ventricular systolic and diastolic function normalized to age, gender and body surface area from steady-state free precession cardiovascular magnetic resonance. Eur Heart J. 2006;27:2879–2888. doi: 10.1093/eurheartj/ehl336. [DOI] [PubMed] [Google Scholar]
- 2.Park J.J., Mebazaa A., Hwang I., Park J., Park J., Cho G. Phenotyping heart failure according to the longitudinal ejection fraction change: myocardial strain, predictors, and outcomes. JAHA. 2020;9 doi: 10.1161/JAHA.119.015009. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Sengeløv M., Jørgensen P.G., Jensen J.S., Bruun N.E., Olsen F.J., Fritz-Hansen T., et al. Global longitudinal strain is a superior predictor of all-cause mortality in heart failure with reduced ejection fraction. JACC: Cardiovasc Imaging. 2015;8:1351–1359. doi: 10.1016/j.jcmg.2015.07.013. [DOI] [PubMed] [Google Scholar]
- 4.Kearney L.G., Lu K., Ord M., Patel S.K., Profitis K., Matalanis G., et al. Global longitudinal strain is a strong independent predictor of all-cause mortality in patients with aortic stenosis. Eur Heart J Cardiovasc Imaging. 2012;13:827–833. doi: 10.1093/ehjci/jes115. [DOI] [PubMed] [Google Scholar]
- 5.Orwat S., Diller G.-P., Kempny A., Radke R., Peters B., Kühne T., et al. Myocardial deformation parameters predict outcome in patients with repaired tetralogy of Fallot. Heart. 2016;102:209–215. doi: 10.1136/heartjnl-2015-308569. [DOI] [PubMed] [Google Scholar]
- 6.Eitel I., Stiermaier T., Lange T., Rommel K.-P., Koschalka A., Kowallick J.T., et al. Cardiac magnetic resonance myocardial feature tracking for optimized prediction of cardiovascular events following myocardial infarction. JACC Cardiovasc Imaging. 2018;11:1433–1444. doi: 10.1016/j.jcmg.2017.11.034. [DOI] [PubMed] [Google Scholar]
- 7.Rajiah P.S., Kalisz K., Broncano J., Goerne H., Collins J.D., François C.J., et al. Myocardial strain evaluation with cardiovascular MRI: physics, principles, and clinical applications. RadioGraphics. 2022;42:968–990. doi: 10.1148/rg.210174. [DOI] [PubMed] [Google Scholar]
- 8.Liu B., Dardeer A.M., Moody W.E., Edwards N.C., Hudsmith L.E., Steeds R.P. Normal values for myocardial deformation within the right heart measured by feature-tracking cardiovascular magnetic resonance imaging. Int J Cardiol. 2018;252:220–223. doi: 10.1016/j.ijcard.2017.10.106. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Kawel-Boehm N., Hetzel S.J., Ambale-Venkatesh B., Captur G., Francois C.J., Jerosch-Herold M., et al. Reference ranges (“normal values”) for cardiovascular magnetic resonance (CMR) in adults and children: 2020 update. J Cardiovasc Magn Reson. 2020;22:87. doi: 10.1186/s12968-020-00683-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Maceira A.M., Tuset‐Sanchis L., López‐Garrido M., San Andres M., López‐Lereu M.P., Monmeneu J.V., et al. Feasibility and reproducibility of feature-tracking-based strain and strain rate measures of the left ventricle in different diseases and genders. Magn Reson Imaging. 2018;47:1415–1425. doi: 10.1002/jmri.25894. [DOI] [PubMed] [Google Scholar]
- 11.Vo H.Q., Marwick T.H., Negishi K. MRI-derived myocardial strain measures in normal subjects. JACC Cardiovasc Imaging. 2018;11:196–205. doi: 10.1016/j.jcmg.2016.12.025. [DOI] [PubMed] [Google Scholar]
- 12.Qu Y.-Y., Li H., Rottbauer W., Ma G.-S., Buckert D., Rasche V. Right ventricular free wall longitudinal strain and strain rate quantification with cardiovascular magnetic resonance based tissue tracking. Int J Cardiovasc Imaging. 2020;36:1985–1996. doi: 10.1007/s10554-020-01895-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.An Investigation of Coronary Heart Disease in Families: The Framingham Offspring Study. Am J Epidemiol. 2017; 185: 1093–1102 doi: 10.1093/aje/kwx110. [DOI] [PubMed]
- 14.Wilson P.W.F., D’Agostino R.B., Levy D., Belanger A.M., Silbershatz H., Kannel W.B. Prediction of coronary heart disease using risk factor categories. Circulation. 1998;97:1837–1847. doi: 10.1161/01.CIR.97.18.1837. [DOI] [PubMed] [Google Scholar]
- 15.Tsao C.W., Gona P.N., Salton C.J., Chuang M.L., Levy D., Manning W.J., et al. Left ventricular structure and risk of cardiovascular events: a framingham heart study cardiac magnetic resonance study. JAHA. 2015;4 doi: 10.1161/JAHA.115.002188. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Salton C.J., Chuang M.L., O’Donnell C.J., Kupka M.J., Larson M.G., Kissinger K.V., et al. Gender differences and normal left ventricular anatomy in an adult population free of hypertension. J Am Coll Cardiol. 2002;39:1055–1060. doi: 10.1016/S0735-1097(02)01712-6. [DOI] [PubMed] [Google Scholar]
- 17.Whelton P.K., Carey R.M., Aronow W.S., Casey D.E., Collins K.J., Dennison Himmelfarb C., et al. 2017 ACC/AHA/AAPA/ABC/ACPM/AGS/APhA/ASH/ASPC/NMA/PCNA guideline for the prevention, detection, evaluation, and management of high blood pressure in adults: a report of the American College of Cardiology/American Heart Association Task Force on Clinical Practice Guidelines. Hypertension. 2018;71 doi: 10.1161/HYP.0000000000000065. [DOI] [PubMed] [Google Scholar]
- 18.Yang W., Xu J., Zhu L., Zhang Q., Wang Y., Zhao S., et al. Myocardial strain measurements derived from MR feature-tracking. JACC Cardiovasc Imaging. 2024;17:364–379. doi: 10.1016/j.jcmg.2023.05.019. [DOI] [PubMed] [Google Scholar]
- 19.Andre F., Steen H., Matheis P., Westkott M., Breuninger K., Sander Y., et al. Age- and gender-related normal left ventricular deformation assessed by cardiovascular magnetic resonance feature tracking. J Cardiovasc Magn Reson. 2015;17:25. doi: 10.1186/s12968-015-0123-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Augustine D., Lewandowski A.J., Lazdam M., Rai A., Francis J., Myerson S., et al. Global and regional left ventricular myocardial deformation measures by magnetic resonance feature tracking in healthy volunteers: comparison with tagging and relevance of gender. J Cardiovasc Magn Reson. 2013;15:8. doi: 10.1186/1532-429X-15-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Taylor R.J., Moody W.E., Umar F., Edwards N.C., Taylor T.J., Stegemann B., et al. Myocardial strain measurement with feature-tracking cardiovascular magnetic resonance: normal values. Eur Heart J Cardiovasc Imaging. 2015;16:871–881. doi: 10.1093/ehjci/jev006. [DOI] [PubMed] [Google Scholar]
- 22.Li G., Zhang Z., Gao Y., Zhu C., Zhou S., Cao L., et al. Age- and sex-specific reference values of biventricular strain and strain rate derived from a large cohort of healthy Chinese adults: a cardiovascular magnetic resonance feature tracking study. J Cardiovasc Magn Reson. 2022;24:63. doi: 10.1186/s12968-022-00881-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Peng J., Zhao X., Zhao L., Fan Z., Wang Z., Chen H., et al. Normal values of myocardial deformation assessed by cardiovascular magnetic resonance feature tracking in a healthy Chinese population: a multicenter study. Front Physiol. 2018;9:1181. doi: 10.3389/fphys.2018.01181. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Zhang Z., Ma Q., Cao L., Zhao Z., Zhao J., Lu Q., et al. Correlation between left ventricular myocardial strain and left ventricular geometry in healthy adults: a cardiovascular magnetic resonance-feature tracking study. Int J Cardiovasc Imaging. 2019;35:2057–2065. doi: 10.1007/s10554-019-01644-3. [DOI] [PubMed] [Google Scholar]
- 25.Cai J., Bryant J.A., Le T.-T., Su B., De Marvao A., O’Regan D.P., et al. Fractal analysis of left ventricular trabeculations is associated with impaired myocardial deformation in healthy Chinese. J Cardiovasc Magn Reson. 2016;19:102. doi: 10.1186/s12968-017-0413-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Weise Valdés E., Barth P., Piran M., Laser K.T., Burchert W., Körperich H. Left-ventricular reference myocardial strain assessed by cardiovascular magnetic resonance feature tracking and fSENC—impact of temporal resolution and cardiac muscle mass. Front Cardiovasc Med. 2021;8 doi: 10.3389/fcvm.2021.764496. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Hwang J., Cha M.J., Kim S.M., Kim Y., Choe Y.H. Relationship between cardiovascular risk factors and myocardial strain values of both ventricles in asymptomatic Asian subjects: measurement using cardiovascular magnetic resonance tissue tracking. Int J Cardiovasc Imaging. 2018;34:1949–1957. doi: 10.1007/s10554-018-1414-1. [DOI] [PubMed] [Google Scholar]
- 28.Liu B., Dardeer A.M., Moody W.E., Hayer M.K., Baig S., Price A.M., et al. Reference ranges for three-dimensional feature tracking cardiac magnetic resonance: comparison with two-dimensional methodology and relevance of age and gender. Int J Cardiovasc Imaging. 2018;34:761–775. doi: 10.1007/s10554-017-1277-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Liu T., Wang C., Li S., Zhao Y., Li P. Age- and gender-related normal references of right ventricular strain values by tissue tracking cardiac magnetic resonance: results from a Chinese population. Quant Imaging Med Surg. 2019;9:1441–1450. doi: 10.21037/qims.2019.08.13. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Truong V.T., Safdar K.S., Kalra D.K., Gao X., Ambach S., Taylor M.D., et al. Cardiac magnetic resonance tissue tracking in right ventricle: feasibility and normal values. Magn Reson Imaging. 2017;38:189–195. doi: 10.1016/j.mri.2017.01.007. [DOI] [PubMed] [Google Scholar]
- 31.Pryds K., Larsen A.H., Hansen M.S., Grøndal A.Y.K., Tougaard R.S., Hansson N.H., et al. Myocardial strain assessed by feature tracking cardiac magnetic resonance in patients with a variety of cardiovascular diseases – a comparison with echocardiography. Sci Rep. 2019;9 doi: 10.1038/s41598-019-47775-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Ananthapadmanabhan S., Vo G., Nguyen T., Dimitri H., Otton J. Direct comparison of multilayer left ventricular global longitudinal strain using CMR feature tracking and speckle tracking echocardiography. BMC Cardiovasc Disord. 2021;21:107. doi: 10.1186/s12872-021-01916-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Taha K., Bourfiss M., Te Riele A.S.J.M., Cramer M.-J.M., Van Der Heijden J.F., Asselbergs F.W., et al. A head-to-head comparison of speckle tracking echocardiography and feature tracking cardiovascular magnetic resonance imaging in right ventricular deformation. Eur Heart J Cardiovasc Imaging. 2021;22:950–958. doi: 10.1093/ehjci/jeaa088. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Raina A., Vaidya A., Gertz Z.M., Susan Chambers, Forfia P.R. Marked changes in right ventricular contractile pattern after cardiothoracic surgery: implications for post-surgical assessment of right ventricular function. J Heart Lung Transplant. 2013;32:777–783. doi: 10.1016/j.healun.2013.05.004. [DOI] [PubMed] [Google Scholar]
- 35.Pettersen E., Helle-Valle T., Edvardsen T., Lindberg H., Smith H.-J., Smevik B., et al. Contraction pattern of the systemic right ventricle. J Am Coll Cardiol. 2007;49:2450–2456. doi: 10.1016/j.jacc.2007.02.062. [DOI] [PubMed] [Google Scholar]
- 36.Kind T., Mauritz G.-J., Marcus J.T., Van De Veerdonk M., Westerhof N., Vonk-Noordegraaf A. Right ventricular ejection fraction is better reflected by transverse rather than longitudinal wall motion in pulmonary hypertension. J Cardiovasc Magn Reson. 2010;12:35. doi: 10.1186/1532-429X-12-35. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Barison A., Ceolin R., Palmieri A., Tamborrino P.P., Todiere G., Grigoratos C., et al. Biventricular tissue tracking with cardiovascular magnetic resonance: reference values of left- and right-ventricular strain. Diagnostics. 2023;13:2912. doi: 10.3390/diagnostics13182912. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Upadhya B., Kitzman D.W. Heart failure with preserved ejection fraction in older adults. Heart Fail Clin. 2017;13:485–502. doi: 10.1016/j.hfc.2017.02.005. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Cheng S., Fernandes V.R.S., Bluemke D.A., McClelland R.L., Kronmal R.A., Lima J.A.C. Age-related left ventricular remodeling and associated risk for cardiovascular outcomes: the multi-ethnic study of atherosclerosis. Circ Cardiovasc Imaging. 2009;2:191–198. doi: 10.1161/CIRCIMAGING.108.819938. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Hung C.-L., Gonçalves A., Shah A.M., Cheng S., Kitzman D., Solomon S.D. Age- and sex-related influences on left ventricular mechanics in elderly individuals free of prevalent heart failure: the ARIC study (atherosclerosis risk in communities) Circ Cardiovasc Imaging. 2017;10 doi: 10.1161/CIRCIMAGING.116.004510. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supplementary material