Author + information
- Received February 13, 2017
- Revision received April 7, 2017
- Accepted April 8, 2017
- Published online July 19, 2017.
- Ahmed Abdi-Ali, MDa,b,
- Robert J.H. Miller, MDa,b,
- Danielle Southern, MScc,
- Mei Zhang, MScc,
- Yoko Mikami, MD, PhDa,
- Merril Knudtson, MDb,c,
- Bobak Heydari, MD, MPHa,b,
- Andrew G. Howarth, MD, PhDa,b,
- Carmen P. Lydell, MDa,d,
- Mathew T. James, MDb,c,
- Stephen B. Wilton, MD, MScb,c and
- James A. White, MDa,b,d,∗ ()
- aStephenson Cardiac Imaging Centre, Libin Cardiovascular Institute of Alberta, University of Calgary, Calgary, Alberta, Canada
- bDepartment of Cardiac Sciences, University of Calgary, Calgary, Alberta, Canada
- cO’Brien Institute for Public Health, University of Calgary, Calgary, Alberta, Canada
- dDepartment of Radiology, University of Calgary, Calgary, Alberta, Canada
- ↵∗Address for correspondence:
Dr. James A. White, Stephenson Cardiac Imaging Centre, Libin Cardiovascular Institute of Alberta, University of Calgary, Suite 0700, SSB, Foothills Medical Centre, 700-1403 29 ST NW, Calgary, Alberta T2N 2T9, Canada.
Objectives The goal of this study was to assess associations between left ventricular (LV) mass, all-cause mortality, and need for revascularization in patients undergoing coronary angiography.
Background LV hypertrophy is associated with adverse cardiovascular outcomes in healthy subjects. However, its influence in those with known or suspected coronary artery disease is poorly understood.
Methods A total of 3,754 patients (mean age 59.3 ± 13.1 years) undergoing invasive coronary angiography and cardiac magnetic resonance (CMR) (mean interval 1.0 ± 1.5 months) were studied. LV mass and volumes were determined from cine images and indexed to body surface area. Analyses were adjusted for CMR variables, medical comorbidities, and severity of coronary artery disease (Duke Jeopardy Score) and were stratified to LV function.
Results At a median of 44.9 months, 315 patients (8.4%) died and 168 patients (4.5%) underwent revascularization. Multivariable analysis showed that each 10 g/m2 increase in LV mass index was associated with a 6% greater risk of mortality (hazard ratio: 1.06; 95% confidence interval [CI]: 1.01 to 1.11; p = 0.02) and a 10% greater need for revascularization (hazard ratio: 1.10; 95% CI: 1.04 to 1.17; p < 0.01). According to pre-defined thresholds, moderate to severe hypertrophy was associated with a 1.7-fold risk of mortality (95% CI: 1.2 to 2.3) and 1.8-fold need for revascularization (95% CI: 1.18 to 2.67). These findings were predominantly observed in those with a left ventricular ejection fraction >35% with respective hazard ratios of 2.93 (95% CI: 1.92 to 4.47) and 2.20 (95% CI: 1.21 to 3.98).
Conclusions LV mass index is an independent predictor of all-cause mortality and need for revascularization. This finding establishes relevance for LV mass measurements in clinical decision-making surrounding both the need and timing of revascularization in this population.
Left ventricular (LV) mass is a recognized marker of cardiovascular risk (1–4). In the Framingham Heart Study, left ventricular hypertrophy (LVH) was an independent predictor of cardiovascular mortality and morbidity in subjects without coronary artery disease (CAD) after adjustment for hypertension (5,6). Among patients with established CAD, limited data have emerged. Although the Heart and Soul Study showed that LVH was associated with sudden cardiac death in this population, this study used 2-dimensional (2D) echocardiographic estimates of LV mass (7). Such estimates are recognized to be less reproducible than those by cardiac magnetic resonance (CMR) (8,9) due to a reliance on geometrical assumptions that may not be valid in some patients (10). Three-dimensional echocardiographic techniques have overcome this limitation and provide estimates very similar to those available from CMR (11); however, prognostic studies have yet to be reported using this technique.
The role for CMR in patients with CAD is increasing due to expanding evidence for robust risk stratification of ischemic and arrhythmic events (12,13). Despite important insights provided by MESA (Multi-Ethnic Study of Atherosclerosis), which performed CMR in healthy subjects, very little is known regarding the prognostic role of LV mass in those with cardiovascular disease (14). A single study, reported by Krittayaphong et al. (15), calculated the left ventricular mass index (LVMI) from gradient echo (GRE) cine images in patients referred for CMR and found associations with major cardiovascular events. However, only 26% of this referral population underwent coronary angiography, limiting any exploration of interactions between CAD burden, cardiovascular events, and LV mass. Accordingly, the relevance of LV mass in the clinical decision-making of patients with CAD remains poorly understood.
To address this knowledge gap, we used a large, well-defined cohort of patients undergoing diagnostic coronary angiography for known or suspected CAD and studied associations between LVMI, as measured by contemporary (steady-state free precession [SSFP]) cine CMR, and the primary outcome of all-cause mortality. The secondary outcome measure of need for revascularization was incrementally studied. Adjustments were performed for both CAD burden (using the Duke Jeopardy Score) and severity of LV dysfunction (using categories of left ventricular ejection fraction [LVEF]).
The patient population was identified from the APPROACH (Alberta Provincial Project for Outcomes Assessment in Coronary Heart Disease) registry, a prospective registry of patients undergoing diagnostic coronary angiography. Database and data collection methods have been previously described in detail (16). In brief, the data collected include demographic characteristics, cardiac risk factors, major clinical comorbidities, previous cardiac events, indication for angiography, coronary anatomy (segmental coding), cardiac medications at the time of angiography, and future clinical outcomes. For the present analysis, we included patients enrolled in the registry between April 1, 2005, and March 30, 2013, with a minimum of 24 months of clinical follow-up and having a CMR examination performed within 3 months of their index angiogram.
All patients provide informed consent at the time of enrollment into the APPROACH registry for their health information to be used for research purposes. Medical comorbidities, including hypertension, are determined at the time of enrollment into the APPROACH registry and represent the attending physician’s interpretation of medical comorbidities, as described in the medical record. This study was approved by the Conjoint Health Region Ethics Board at the University of Calgary.
CMR imaging and analysis
All CMR examinations were performed by using a 1.5-T system (MAGNETOM Avanto, Siemens Healthcare, Erlangen, Germany). Standard cine SSFP imaging was performed for the evaluation of LV volume, function, and mass. Quantitative image analysis was performed by experienced readers using commercially available software (cvi42, Circle Cardiovascular Imaging Inc., Calgary, Alberta, Canada). Cine images underwent semi-automated contour tracing to obtain the left ventricular end-diastolic volume (LVEDV), LV end-systolic volume, LVEF, and LV mass. The latter was determined at end-diastole without the inclusion of papillary muscles. All volume and mass measures were indexed to body surface area.
Using previously published CMR reference values, categories of LVH were defined a priori relative to age and sex-matched control subjects; mild LVH was defined as >2 SDs and moderate to severe LVH as >4 SDs above respective mean reference values, as shown in Online Table 1 (17). Although several SSFP-based studies exist describing similar normal reference values for LV mass (18), for maximal translatability to published cohorts, we chose to apply reference values from the widely cited study by Hudsmith et al. (17).
To explore the influence of LVH pattern, patients’ LV mass pattern was classified as follows: normal (n = 2,525), concentric remodeling (n = 141), eccentric hypertrophy (n = 884), and concentric hypertrophy (n = 204). Concentric remodeling was defined as normal LVMI with an increase in concentricity (LVM/LVEDV); eccentric LVH was defined as increased LVMI without an increase in concentricity; and concentric LVH was defined as an increase in both values. LV dilatation was defined as >95th percentile LVEDV index (>102 ml/m2 for women and >108 ml/m2 for men) (17), with a similar definition for increased concentricity (>0.95 g/ml for women and >1.15 g/ml for men) (19). All measurements were performed without any knowledge of patient outcomes.
The primary outcome was all-cause mortality, determined by linkage with Alberta Vital Statistics. The secondary outcome was the symptom-driven need for percutaneous or surgical coronary revascularization at least 30 days after the index angiogram, ascertained from APPROACH. A 30-day “blanking period” was applied after the index coronary angiogram to minimize the influence of revascularization related to baseline angiographic findings as well as procedural complications.
Baseline clinical data are reported as mean ± SD for continuous variables or as percentages for categorical data. Differences between the 2 groups were analyzed using the Wilcoxon rank sum test for continuous variables and the Fisher exact test for categorical variables.
LVMI was considered both as a continuous variable and using the normal, mild hypertrophy and moderate/severe hypertrophy categories defined earlier. Event-free survival, stratified according to categories of LVMI, was estimated using Kaplan-Meier analysis, and significance was assessed by using the log-rank method.
We investigated associations between LVMI categories and primary and secondary outcomes Cox proportional hazards models, adjusting for all variables in Table 1, including baseline clinical characteristics, comorbidities, CMR findings, and the severity of CAD as determined at the time of angiography quantified by using the Duke Jeopardy Score (20). A stepwise backward elimination method was used to refine the final models, excluding variables with the worst association until only variables signiﬁcantly associated with the outcome remained (p < 0.05) (21). Similar analyses were conducted with a forward stepwise selection method and with a model including all variables with similar findings; these results are outlined in Online Table 2. The proportional hazards assumption was confirmed by using Schoenfeld residuals. We performed a priori stratification of analyses according to LV systolic function by using the following 3 groups: LVEF >50%; LVEF 35% to 50%; and LVEF <35% (22). The associations of LVH, LV dilatation, increased concentricity, and LVH pattern were also investigated in univariate and multivariate models. We assessed for interactions between LVMI and all other variables in the final multivariable models. There were no significant interactions according to sex. Collinearity was assessed by using covariance matrix estimates, with no substantial collinearity identified in the final models (absolute values all <0.50). All analyses were performed by using Stata version 13 (StataCorp, College Station, Texas).
A total of 3,754 patients met the study inclusion criteria. Online Figure 1 outlines patient inclusion details. Diagnostic angiography and CMR were performed with a mean interprocedural interval of 1.0 ± 1.5 months.
Baseline clinical characteristics
Table 1 presents the baseline characteristics of the study population. The mean age of the population was 59.3 ± 13.1 years (interquartile range [IQR]: 50.9 to 68.9 years) with 1,040 (27.7%) being female. The median age of the population was 60.1 years. The mean LVEF was 45.7 ± 16.3% (IQR: 33% to 59%) with 984 (26.2%) patients having an LVEF ≤35%. The mean LVMI of the population was 69.5 ± 23.4 g/m2 (IQR: 53.4 to 81.0 g/m2). A total of 1,700 patients (59.3%) had evidence of late gadolinium enhancement (LGE) consistent with previous myocardial injury. At 30 days from index angiography, 854 (22.8%) patients had undergone a revascularization procedure.
Primary outcome: all-cause mortality
Over a median follow-up of 44.9 months (IQR: 24.6 to 76.6 months), 315 (8.4%) patients died. Significant differences were observed in a number of baseline clinical variables: patients who died were older (67 years vs. 59 years; p < 0.01), had lower systolic blood pressure (118 mm Hg vs. 121 mm Hg; p < 0.01), and had more medical comorbidities (Table 1). With respect to baseline CMR variables, patients who died had a lower LVEF (37.9 ± 16.4% vs. 45.7 ± 16.3%; p < 0.01), higher LVEDV index (123.5 ± 55.4 ml/m2 vs. 109.1 ± 42.6 ml/m2; p < 0.01), higher LVMI (73.8 ± 23.4 g/m2 vs. 69.5 ± 23.4 g/m2; p < 0.01), and were more likely to have LGE (56.3% vs. 44.3%; p < 0.01).
Univariable analysis revealed that both LVEF and LVMI were strongly associated with the primary outcome. Each 10% reduction in LVEF was associated with a 1.4-fold increased risk of mortality. The pre-specified LVEF threshold of <35% provided a 2.4-fold increased risk of mortality (p < 0.01). By comparison, each 10 g/m2 increase in the LVMI was associated with a 1.1-fold increased risk of mortality (p < 0.01).
LV mass was analyzed categorically according to mild (>2 SDs above reference mean value for sex) and moderate to severe (>4 SDs above reference mean value for sex) LVH subgroups. Using these criteria, 638 (18.2%) patients exhibited mild LVH and 450 (12.9%) exhibited moderate to severe LVH. Relative to those with normal range LV mass, patients with mild LVH had an unadjusted hazard ratio (HR) of 1.79 (95% confidence interval [CI]: 1.36 to 2.37; p < 0.01) for mortality, and patients with moderate to severe LVH had an unadjusted HR of 2.43 (95% CI: 1.81 to 3.26; p < 0.01) for mortality. LV mass was also analyzed according to study population quartiles (first quartile used as reference) and revealed the following HRs for mortality: Q2 (62 to 76 g/m2) HR: 0.89 (95% CI: 0.63 to 1.25); Q3 (77 to 95 g/m2) HR: 1.37 (95% CI: 0.99 to 1.88); and Q4 (96 to 244 g/m2) HR: 1.67 (95% CI: 1.22 to 2.29). The pooled HR for patients with an LVMI in Q3 or Q4 (77 to 244 g/m2) was 1.61 (95% CI: 1.24 to 2.09). Our results were consistent in patients with a history of myocardial infarction, stroke, or heart failure as well as in patients without such history.
Kaplan-Meier event-free survival estimates were performed stratified according to categories of LVH severity and LVMI population quartiles. The results of these analyses are shown in Figures 1A and 1B. The cumulative mortality rate in those with moderate to severe LVH was 12.9% versus 10.5% in those with mild LVH and 7.2% in those without LVH (p < 0.01). Similarly, cumulative mortality was higher with increasing quartiles of LVMI (p < 0.01).
Multivariable analysis, performed inclusive of all relevant baseline clinical and CMR variables (inclusive of age, LVEF <35%, presence of LGE, and severity of CAD represented by the Duke Jeopardy Score), revealed several independent predictors of the primary outcome (model 1, Table 2). Of these, LVMI emerged as a strong, independent predictor of mortality with an adjusted HR of 1.06 per 10 g/m2 (95% CI: 1.01 to 1.11). When entered as a categorical variable, moderate to severe LVH exhibited an adjusted HR of 1.71 (95% CI: 1.25 to 2.34), whereas an LVMI above the median value for the whole study population (>77 g/m2) showed an adjusted HR of 1.11 (95% CI: 0.83 to 1.49) (models 2 and 3, Table 2).
LVEF was found to interact with LVMI in the final model. However, repeat multivariable analyses performed after exclusion of LVEF showed similar findings. An expanded investigation of LVEF and its influence on associations between LVMI and all-cause mortality are described later in the Results.
Secondary outcome: need for revascularization
During follow-up, 168 (4.5%) patients were admitted to the hospital for revascularization (54 for percutaneous coronary intervention and 114 for coronary artery bypass grafting).
Univariable analyses demonstrated LVMI provided an HR of 1.13 per 10 g/m2 (95% CI: 1.07 to 1.19; p < 0.01) for occurrence of the secondary outcome. An incremental rise in the need for revascularization was identified for increasing quartiles of LVMI, ranging from an unadjusted HR of 1.00 in Q1 (reference) to 3.17 in Q4 (95% CI: 1.91 to 5.27; p < 0.01). Similarly, mild LVH was associated with a 1.7-fold increase in risk (unadjusted HR: 1.67; 95% CI: 1.14 to 2.46; p < 0.01), and moderate to severe LVH provided a 2.3-fold increase in risk (unadjusted HR: 2.28; 95% CI: 1.52 to 3.42; p < 0.01).
Multivariate analysis showed that LVMI remained independently predictive of revascularization, with an HR of 1.10 per 10 g/m2 (95% CI: 1.04 to 1.17; p < 0.01) after adjustment for all eligible baseline covariates (Table 3). Categorical analyses revealed that moderate to severe LVH provided a 1.8-fold increased risk (adjusted HR: 1.77; 95% CI: 1.18 to 2.67; p < 0.01), whereas a nonsignificant trend was identified for those with mild LVH (adjusted HR: 1.20; 95% CI: 0.81 to 1.77; p = 0.36). Kaplan-Meier estimates for freedom from revascularization, stratified according to quartiles and LVH severity, are shown in Figures 2A and 2B. Similar curves are also provided for the composite secondary outcome of revascularization or mortality (Figures 3A and 3B). There were no significant interactions with LVMI, and our findings remained significant after excluding events that occurred >30 days after angiography but before CMR.
Influence of LV systolic dysfunction
An a priori subgroup analysis was performed to identify the relative predictive utility of LVMI across 3 ranges of systolic function: severe (LVEF <35%), mild to moderate (LVEF 35% to 50%), and normal (LVEF ≥50%). The number of patients in each subgroup was 984 (26.2%), 929 (24.7%), and 1,841 (49.0%), respectively. The baseline characteristics of these subgroups are shown in Online Table 3. A statistically significant interaction was observed between LV dysfunction category and LVMI (p = 0.02). Kaplan-Meier event-free survival estimates stratified according to LVH severity are shown for patients with LVEF ≤35% and those with LVEF >35% in Figures 4A and 4B.
Overall, patients with severe LV dysfunction were more likely to experience mortality versus other subgroups with a cumulative event rate of 13.6% versus 6.6% (p < 0.01). Results of multivariable analysis for all-cause mortality in each respective subgroup are shown in Online Table 4.
Among patients with LVEF >35%, moderate to severe LVH was associated with a 2.9-fold increased risk of mortality (adjusted HR: 2.93; 95% CI: 1.92 to 4.48; p < 0.01), whereas mild LVH did not achieve a statistically significant association. However, among patients with severe LV dysfunction, no significant associations between LVMI and the primary outcome were identified.
Stratified analysis was also performed for revascularization, and the results are shown in Online Table 5. In patients with LVEF >35%, moderate to severe LVH was associated with a 2.2-fold increase in the need for revascularization (adjusted HR: 2.20; 95% CI: 1.21 to 3.98; p < 0.01). Meanwhile, we found no significant associations in patients with severe LV dysfunction.
Influence of LVH pattern
In univariate analysis, the presence of increased LVMI (unadjusted HR: 2.03; 95% CI: 1.63 to 2.56; p < 0.01), concentricity (unadjusted HR: 1.79; 95% CI: 1.28 to 2.50; p < 0.01), and LV dilatation (unadjusted HR: 1.59; 95% CI: 1.26 to 2.00; p < 0.01) were associated with an increased risk of all-cause death. When pattern of LVH was assessed, eccentric hypertrophy (unadjusted HR: 1.94; 95% CI: 1.51 to 2.49; p < 0.01) and concentric hypertrophy (unadjusted HR: 2.66; 95% CI: 1.80 to 3.91; p < 0.01) were associated with significant risk of mortality, whereas concentric remodeling (unadjusted HR: 1.47; 95% CI: 0.78 to 2.79; p = 0.23) was not. Unadjusted Kaplan-Meier survival curves are shown in Online Figure 2.
After multivariate adjustment for all variables as described earlier, the presence of increased LVMI (adjusted HR: 1.33; 95% CI: 1.03 to 1.71; p = 0.03) and concentricity (adjusted HR: 2.55; 95% CI: 1.79 to 3.63; p < 0.01) continued to be independently associated with all-cause mortality, whereas LV dilatation (adjusted HR: 0.84; 95% CI: 0.62 to 1.13; p = 0.25) was not. In a similar analysis, concentric hypertrophy was independently associated with all-cause mortality (adjusted HR: 3.03; 95% CI: 2.04 to 4.48; p < 0.01), whereas no independent association was found with eccentric hypertrophy (adjusted HR: 1.08; 95% CI: 0.80 to 1.40; p = 0.69) and concentric remodeling (adjusted HR: 1.74; 95% CI: 0.90 to 3.34; p = 0.10).
In this large cohort of well-characterized patients undergoing coronary angiography, elevations in LVMI were associated with all-cause mortality after adjustment for all baseline covariates, inclusive of CAD burden. Mortality progressively worsened with increasing severity of LVH, most notably among patients with normal or mild to moderately depressed LV function, subgroups comprising three-quarters of this referral population. We incrementally identified similar and independent associations between LVMI and need for revascularization. Interestingly, the presence of moderate to severe LVH was associated with a similarly increased risk as the presence of LV dysfunction and diabetes. Overall, these findings confirm a significant association between LVMI and future cardiovascular events in patients referred for diagnostic coronary angiography.
Several previous studies have aimed to describe the adverse influence of LVH on future cardiovascular outcomes by using available surrogate markers. Although inexpensive, ECG-based screening methods for LVH provide poor sensitivity for the detection of elevated LV mass (23). Limitations have also been identified, with 2D echocardiographic estimates of LV mass showing reduced precision versus the reference standard of CMR (8,9). Despite these limitations, the presence of LVH by such surrogates has provided sufficient support to establish widespread acceptance of LVH being a predictor of major cardiovascular events in otherwise healthy subjects (2,5,6,24-26). However, only 2 studies have previously attempted to develop similar linkages in patients with cardiovascular disease by using echocardiography-based techniques (7,26). The first of these, published by Quiñones et al. (26), was an analysis of 2D echocardiograms performed on patients in the SOLVD (Studies of Left Ventricular Dysfunction) clinical trials. Although not representing contemporary care, this important study identified that LV mass estimates were associated with the composite endpoint of all-cause mortality and cardiovascular hospitalization. The second study, published by Turakhia et al. (7), described findings from the Heart and Soul Study linking echocardiography-based estimates of LVMI to mortality in patients with stable CAD. Using a threshold value of LVMI >95 g/m2 in women and >115 g/m2 in men, elevated LVMI was associated with a 2-fold increase in all-cause mortality and a 3-fold increase in sudden cardiac death. This study did not describe influences of LV mass on future revascularization and did not adjust for differences in the severity of CAD on future outcomes.
It is important to recognize that although CMR demonstrates improved reproducibility versus conventional echocardiographic techniques for the estimation of LV mass (8,9), this measure does not reflect the accuracy of a technique. To date, there have been 2 echocardiographic studies and 1 CMR-based study comparing imaging-based LV mass versus those obtained at autopsy. The first study compared pre-mortem (in vivo) M-mode echocardiography-based estimates versus post-mortem LV mass in 34 subjects; it was published in 1977 by Devereux and Reichek (27). This landmark study described strong correlations (r = 0.86 to 0.96) using a variety of formulas. Upon clinical migration to 2D echocardiography techniques, a similar study was reported, demonstrating more modest agreement using both the area length and truncated ellipsoid methods (r = 0.72 and 0.71) (28). Most recently, a study using CMR reported high correlation (r = 0.95) to autopsy-based measurement; however, it used post-mortem imaging (29). Accordingly, a comparison of in vivo CMR versus post-mortem LV mass has yet to be described.
MESA (Multi-Ethnic Study of Atherosclerosis) enrolled 6,814 adults age 45 to 84 years old who were free from overt cardiovascular disease in 6 U.S. communities and obtained baseline CMR imaging by using a standardized protocol. An analysis of this cohort showed that LVH (defined by LVM/height ≥95th percentile) was associated with increased all-cause mortality (HR: 1.8; 95% CI: 1.1 to 2.9) (14). A study by Krittayaphong et al. (15) described the prognostic value of CMR-based LV mass quantification in a referral population of patients with cardiovascular disease; however, it was not specific to patients with CAD. In this population, a higher occurrence of major adverse events (inclusive of death, nonfatal myocardial infarction, and cardiovascular hospitalization) was observed in those with an LVMI above the 50th percentile (i.e., Q3 and Q4). This group experienced a 2.3-fold (95% CI: 1.3 to 4.0) increase in the rate of events over a median follow-up of 926 ± 582 days (p < 0.001) after adjustment for LVEF and the presence of LGE. Because only one-quarter of this population underwent coronary angiography, the analysis was not intended to address a CAD referral population and did not stratify for CAD severity or degree of LV dysfunction. The influence of LV mass on future revascularization was similarly not described.
The present study provides important insights into the influence of LV mass on future cardiovascular events in patients with known or suspected CAD. As the largest study to date, and of the longest duration of follow-up, it provides robust evidence for increased LV mass contributing to mortality in patients receiving contemporary cardiovascular care. It also allowed for subgroup analysis in patients with normal, mild to moderate, or severe LV systolic dysfunction, demonstrating that its most important influence is among those without severe dysfunction. This observation is congruent with our finding that concentric LVH was associated with worse outcomes, whereas eccentric LVH was not. Patients with eccentric LVH include patients with dilated cardiomyopathy, in whom outcomes are driven primarily by reduced LVEF. Beyond these important contributions, our study also supports logical pathophysiological postulates for LV mass contributing to clinical ischemia burden. This finding is evidenced by an elevated need for revascularization in those patients with higher degrees of hypertrophy after correction for baseline CAD burden. This important finding may likely reflect greater supply and demand mismatch in these patients, supporting LV mass as a potentially important clinical marker to contextualize the relevance of findings on diagnostic angiography.
First, as a retrospective observational cohort study, a causal relationship cannot be established between elevations in LV mass and future cardiovascular events. Second, the administrative data used to determine vital status of this population did not allow for differentiation of cause of mortality; therefore, only all-cause mortality could be reported. Third, we recognize that measurements of LV mass may vary between interpreting physicians and that this study did not use core laboratory measurements or perform reproducibility analyses; rather, this study evaluated LV mass measurements performed in routine clinical practice, and it therefore represents a conservative yet “real-world” estimate of the value provided by LV mass measurements to predict cardiovascular events in this population. In addition, due to the methods of this study, we were unable to determine the duration and severity of hypertension, which may be an important driver of LVH-related outcomes. As a large, retrospective cohort study, this analysis did not include the papillary muscles as part of the LV mass. Although this approach is reasonable, given the lack of objective criteria to discriminate papillary muscles from minor muscle bundles or trabeculation, recent studies have highlighted variable contributions of papillary muscles to global LV mass in patients with different cardiomyopathy states, particularly Fabry cardiomyopathy (30,31). Accordingly, improvements in predictive utility for LV mass may be achieved through consideration of such techniques. Finally, only a small proportion of patients from the APPROACH registry underwent CMR within 3 months of coronary angiography and were included in this analysis. This approach reflects an inherent selection bias related to typical indications for clinical CMR imaging in this patient population. As such, these potential biases must be considered when generalizing to other CAD patient populations, as has been suggested for other observational cohort studies (32,33).
LV mass was a powerful and independent predictor of mortality in patients referred for diagnostic angiography and was associated with a greater need for revascularization. Consideration of this marker for optimal clinical decision-making in patients referred for diagnostic angiography, particularly those with LVEF >35%, is justified and warrants expanded investigation.
COMPETENCY IN MEDICAL KNOWLEDGE: LVMI provides valuable prognostic information regarding the occurrence of death or need for revascularization, particularly in patients with preserved LV function.
COMPETENCY IN PATIENT CARE AND PROCEDURAL SKILLS: The presence of elevated LVMI identifies patients at higher risk for clinical events who may benefit from more aggressive medical therapy.
TRANSLATIONAL OUTLOOK: Although LVMI is a powerful predictor of clinical events, its pathophysiological basis is not completely delineated in this study and requires further investigation.
The authors thank the members of the APPROACH research working group and the staff of the Stephenson Cardiac Imaging Centre.
For supplemental tables and figures, please see the online version of this article.
APPROACH was initially funded with a grant from the W. Garfield Weston Foundation. The ongoing operation of the APPROACH project has been made possible by support from Alberta Health Services (Calgary Zone, Edmonton Zone), Libin Cardiovascular Institute of Alberta, and Mazankowski Alberta Heart Institute. The APPROACH initiative has also received contributions from Alberta Health and Wellness and several industry sponsors (Merck Frosst Canada Inc., Eli Lily Canada Inc., and Servier Canada Inc.) to support the basic infrastructure of this cardiac registry initiative. Dr. White has received funding from the Calgary Health Trust; is supported by an Early Investigator Award from the Heart and Stroke Foundation of Alberta; is Chief Medical Officer of Cohesic Inc., in which he holds shares; and has received grant support from Circle Cardiovascular Inc. Dr. Wilton has received grant support from St. Jude Medical and Alberta Health Services; and has served as a consultant for Boehringer-Ingelheim and Arca Biopharma. Dr. James is supported by a KRESCENT New Investigator Award, a joint initiative of the Kidney Foundation of Canada, the Canadian Institute of Health Research and the Canadian Society of Nephrology, and by a Canadian Institutes of Health Research New Investigator Award. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose. Drs. Abdi-Ali and Miller are joint first authors.
- Abbreviations and Acronyms
- coronary artery disease
- confidence interval
- cardiac magnetic resonance
- gradient echo
- hazard ratio
- interquartile range
- late gadolinium enhancement
- left ventricular
- left ventricular end-diastolic volume
- left ventricular ejection fraction
- left ventricular hypertrophy
- left ventricular mass index
- steady-state free precession
- Received February 13, 2017.
- Revision received April 7, 2017.
- Accepted April 8, 2017.
- 2017 American College of Cardiology Foundation
- Krumholz H.M.,
- Larson M.,
- Levy D.
- Bluemke D.A.,
- Kronmal R.A.,
- Lima J.A.,
- et al.
- Haider A.W.,
- Larson M.G.,
- Benjamin E.J.,
- Levy D.
- Bottini P.B.,
- Carr A.A.,
- Prisant L.M.,
- Flickinger F.W.,
- Allison J.D.,
- Gottdiener J.S.
- Seo H.Y.,
- Lee S.P.,
- Park J.B.,
- et al.
- Marwick T.H.,
- Gillebert T.C.,
- Aurigemma G.,
- et al.
- Greenwood J.P.,
- Herzog B.A.,
- Brown J.M.,
- et al.
- Califf R.M.,
- Phillips H.R. 3rd.,
- Hindman M.C.,
- et al.
- Wang T.J.,
- Evans J.C.,
- Benjamin E.J.,
- Levy D.,
- LeRoy E.C.,
- Vasan R.S.
- Levy D.,
- Labib S.B.,
- Anderson K.M.,
- Christiansen J.C.,
- Kannel W.B.,
- Castelli W.P.
- Chiang B.N.,
- Perlman L.V.,
- Fulton M.,
- Ostrander L.D. Jr..,
- Epstein F.H.
- Kannel W.B.,
- Doyle J.T.,
- McNamara P.M.,
- Quickenton P.,
- Gordon T.
- Quiñones M.A.,
- Greenberg B.H.,
- Kopelen H.A.,
- et al.
- Devereux R.B.,
- Reichek N.
- Farber N.J.,
- Reddy S.T.,
- Doyle M.,
- et al.
- Barbash G.I.,
- Reiner J.,
- White H.D.,
- et al.