Author + information
- Received November 30, 2015
- Revision received January 11, 2016
- Accepted January 14, 2016
- Published online November 1, 2016.
- Sameer Raina, MDa,∗ (, )
- Shelly Y. Lensing, MSb,
- Ramez S. Nairooz, MDa,
- Naga Venkata K. Pothineni, MDa,
- Abdul Hakeem, MDa,c,
- Sabha Bhatti, MDa and
- Tarun Pandey, MDd
- aDepartment of Cardiology, University of Arkansas for Medical Sciences, Little Rock, Arkansas
- bDepartment of Biostatistics, University of Arkansas for Medical Sciences, Little Rock, Arkansas
- cCentral Arkansas VA Health System, Little Rock, Arkansas
- dDepartment of Radiology, University of Arkansas for Medical Sciences, Little Rock, Arkansas
- ↵∗Reprint requests and correspondence:
Dr. Sameer Raina, Division of Cardiovascular Medicine, University of Arkansas for Medical Sciences, 4301 West Markham Street, Little Rock, Arkansas 72205.
Objectives The aim of this study was to access the prognostic implication of late gadolinium enhancement (LGE) in patients with systemic amyloidosis undergoing cardiac magnetic resonance (CMR).
Background Cardiac amyloidosis confers significantly worse prognosis in patients with systemic amyloidosis. CMR imaging has emerged as an attractive noninvasive modality to diagnose cardiac involvement in patients with systemic amyloidosis. We performed a systemic review and meta-analysis to evaluate the prognostic role of LGE-CMR imaging in patients with systemic amyloidosis.
Methods Electronic databases MEDLINE, PubMed, Embase, and Cochrane were systematically searched to identify studies evaluating the association between LGE-CMR and prognosis in systemic amyloidosis with cardiac involvement. The present study was designed to systematically review and assess the association between LGE and the primary endpoint of all-cause mortality. A random effects model was used to calculate a pooled odds ratio using inverse-variance weighting.
Results Data were included from 7 studies with a total of 425 patients and a mean follow-up of 25 months. Patients had a weighted average age of 64 years and left ventricular ejection fraction of 59.2%; 67% were male. Endomyocardial biopsy was positive for amyloidosis in 20%, whereas LGE was present in 73% of patients. LGE-positive patients had increased overall mortality compared with those without LGE (pooled odds ratio: 4.96; 95% confidence interval [CI]: 1.90 to 12.93; p = 0.001). For the LGE group, the pooled death rate was 0.07 (95% CI: 0.03 to 0.19) events per year and for the LGE+ group, the rate was 0.25 (95% CI: 0.16 to 0.39 per year; p = 0.001). The proportion of patients with cardiac biopsy within each study ranged from 3% to 68%, and the relationship between LGE status and death did not vary according to cardiac biopsy proportion across studies.
Conclusions LGE on CMR in patients with systemic amyloidosis with known or suspected cardiac amyloidosis is associated with increased risk of all-cause mortality.
Systemic amyloidosis is a relatively rare multisystem disease caused by the deposition of misfolded protein in various tissues and organs. Cardiac involvement is a leading cause of morbidity and mortality in these patients, especially in primary light-chain (AL) amyloidosis, which constitutes almost 85% of patients with systemic amyloidosis in the United States (1,2). Death is attributed to cardiac involvement from congestive heart failure or arrhythmias in at least one-half of patients with AL (1). The presence of cardiac involvement and its relative predominance varies with the type of amyloidosis. The most common form of leading to cardiac amyloidosis (CA) is the AL type resulting from deposition of amyloid light chains. Amyloid-transthyretin (ATTR), also called as senile amyloidosis and amyloid A protein, also known as secondary amyloidosis, are other, less-common types resulting in CA. Although secondary amyloidosis almost never affects the heart in any clinically significant manner (2,3), AL amyloidosis is frequently responsible for severe cardiac involvement. Regardless of the type of CA, the use of cardiac magnetic resonance (CMR) provides important functional and morphological information in these patients. CMR is similar to echocardiography in evaluation of myocardial function, though the latter is superior for evaluating and quantifying diastolic abnormalities and in delineating morphological abnormalities. One unique advantage of CMR is the ability to study myocardial enhancement in various cardiomyopathies and its ability to differentiate various forms of ischemic and nonischemic cardiomyopathies based on the pattern of enhancement on late gadolinium enhancement (LGE). The appearance of global, subendocardial LGE has been described as highly characteristic of cardiac amyloid (4,5). In multiple studies, LGE-CMR has been found to be the most accurate predictor of endomyocardial biopsy-positive CA, with excellent diagnostic accuracy approaching >90% (5). Although the diagnostic performance of LGE-CMR has been well-validated for CA, its prognostic value has not been well-established, with several studies yielding mixed results. These studies have in general been single center, with small sample populations and few events (5–11). Considering high mortality in patients with systemic amyloidosis who develop cardiac amyloidosis, the use of readily available imaging modalities such as CMR may prove helpful to study possible prognostic markers such as LGE for better risk stratification and management in these patients. A quantitative evaluation and synthesis of this information would hence be essential in elucidating the role of CMR in prognostication of patients with CA. We performed a systematic review and meta-analysis of studies evaluating the prognostic value of LGE-CMR in patients with known or suspected CA.
We included studies assessing patients with systemic amyloidosis with known or suspected CA having had CMR imaging for evaluation of LGE and a minimum follow-up of 12 months. Studies with ischemic and nonischemic cardiomyopathies, acute myocarditis, and hypertrophic and infiltrative cardiomyopathies were excluded.
To identify eligible studies, 2 physician investigators (S.R. and S.Y.L.) independently searched (October 2015) Cochrane CENTRAL, EMBASE, and PubMed for studies assessing prognosis in patients with systemic amyloidosis with known or suspected CA after undergoing CMR for LGE. Key words used were (“prognosis” OR “outcome”) AND (“cardiac amyloidosis” OR “systemic amyloidosis”) AND (“delayed gadolinium enhancement” OR “late gadolinium enhancement” OR “magnetic resonance imaging”). In addition, we explored “related articles” for key publications in PubMed and reviewed citations from eligible studies. There were no language restrictions. We performed our systematic review and meta-analysis in accordance with MOOSE (Meta-analysis Of Observational Studies in Epidemiology) and PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines (12,13).
Studies evaluating the prognostic role of LGE-CMR in patients with known or suspected CA with at least 12 months of follow-up with mortality as the endpoint were included. Other studies evaluating role of LGE-CMR in patients with ischemic and nonischemic cardiomyopathies, acute myocarditis, and hypertrophic and infiltrative cardiomyopathies were excluded. All abstracts and full-text reports of articles meeting selection criteria were retrieved and studied. The PubMed search query on October 20, 2015, produced 115 initial results (Figure 1). Of these, 105 were excluded by title or abstract as unrelated to the scientific question. Ten were retrieved for detailed evaluation. Of the 10 potentially appropriate studies, only 7 evaluated the selected endpoints of all-cause mortality in systemic amyloidosis patients with known or suspected CA using LGE by CMR. These 7 studies were then included in the meta-analysis. Any disagreements were resolved by consensus.
Two physician investigators independently abstracted following demographic data: author, year of publication, study design, sample size, age, percentage male, hypertension, New York Heart Association (NYHA) functional class, left ventricular ejection fraction (LVEF), LGE status, cardiac and extracardiac biopsy, type of amyloidosis (AL, ATTR, other), and follow-up duration. Primary outcome was all-cause mortality. In case of discrepancy, the final decision was made by consensus of all authors. Further clarifications regarding data were made via e-mail with the study author.
A random effects model was used to calculate a pooled odds ratio (OR) using inverse-variance weighting (14). Heterogeneity among studies was assessed using the Q and I2 statistics. A sensitivity analysis to assess the robustness of the results was performed by rerunning the analyses excluding 1 study at a time. Publication bias was evaluated by examining the funnel plot, Egger’s test (15), and Peters’ test (16). The relationship between cardiac biopsy on the results was examined in a bubble plot of the individual study ORs versus the proportion of patients with cardiac biopsy with bubble size weighted by study size. Calculating person-years of follow-up using average or median follow-up times, death rates were calculated for individual studies and pooled using a Poisson-normal mixed model for the meta-analysis of incidence rates (17). Analyses related to pooling ORs were performed using Review Manager (RevMan) 5, version 3, freeware package (The Nordic Cochrane Centre, The Cochrane Collaboration, 2014, Copenhagen, Denmark), and rmeta (Thomas Lumley 2012, Seattle, Washington) and meta (Guido Schwarzer 2015, Freiburg, Germany) packages in R 3.2.0 (R Core Team 2015, R Foundation for Statistical Computing, Vienna, Austria). Meta-analysis of incidence rates was done using SAS version 9.4. Statistical significance for hypothesis testing was set at the 2-sided 0.05 level.
Seven studies with 425 patients formed the final dataset. Table 1 shows the inclusion and exclusion criteria and primary endpoints evaluated in each of these studies. Five of these studies were prospective, single-center studies, whereas the rest were retrospective studies (5–11). All-cause mortality was recorded in all studies. The detection of LGE was done using a 1.5-T (Siemens Medical Solutions, Erlangen, Germany) (5,6,10,11), Philips Medical Systems (Amsterdam, the Netherlands) (7,9), and General Electric (Fairfield, Connecticut) (8) scanner with 0.1 mmol/kg (6,8,11), 0.15 mmol/kg (10), and 0.2 mmol/kg (5,7,9) of gadolinium contrast injection. Most studies used 10 min of delay to detect LGE.
Table 2 shows the individual and pooled characteristics from these studies; some patients within studies were excluded from outcome specific measures due to not having biopsy-proven systemic amyloidosis or having nondiagnostic LGE images. Males composed 67% of the subjects, weighted mean age was 64 years and LVEF was 59.2%. The patients had a median NYHA functional class III heart failure based on 5 studies; the study by Mekinian et al. (7) mentioned presence of heart failure in (59%), no data for NYHA functional class was provided. The mean follow-up period was 2.1 years. Overall prevalence of LGE in these studies was 73% (range 28% to 84%), whereas one-fifth of patients had endomyocardial biopsy–proven CA.
Table 3 shows the CMR sequence, pattern of LGE, and imaging characteristics of individual studies. Results varied in different studies for the association of LGE with prognosis in patients with systemic amyloidosis with known or suspected CA. The study by Maceira et al. (6) showed no difference in survival with respect to presence or absence of LGE (p = 0.359). Similarly, the study by Ruberg et al. (9) showed that LGE is highly sensitive and specific for the identification of cardiac involvement but does not predict survival (p = 0.62). On the other hand, the study by Austin et al. (5) showed that the characteristic delayed hyperenhancement-CMR pattern is more accurate for diagnosis and is a stronger predictor of 1-year mortality in patients with suspected CA (p = 0.03). The study by Mekinian et al. (7) found that the presence of a positive CMR in AL amyloidosis was associated with a significantly increased risk of death (p = 0.01), with three-fourths of these patients being positive for LGE. The study by Migrino et al. (8) demonstrated a higher proportion of all-cause mortality at 1 year in patients positive for LGE on CMR (p = 0.03). White et al. (10) showed that presence of diffuse HE was the most important predictor of death in the group with suspected CA (p < 0.0001). Fontana et al. (11) recently published the largest study on prognosis of LGE in systemic amyloidosis. They also defined the patterns into subendocardial and transmural LGE. Transmural LGE was a significant predictor of mortality in the overall population (p < 0.0001). Most studies showed higher incidence of mortality in patients with heart failure, which varied with severity of heart failure.
LGE imaging characteristics
Considerable variability was noted in the type of sequence used for LGE (Table 3). Phase-sensitive inversion recovery (PSIR), which is regarded as the most robust of all LGE sequences, was used in 3 of 7 studies for determining the LGE (5,8,11). Of these 3 studies, 1 used PSIR in only 43% of the subjects (11). Maceira et al. (6) used intramyocardial T1 difference between subepicardium and subendocardium for diagnosis of CA with 85% accuracy in predicting mortality in systemic AL amyloidosis patients. Mekinian et al. (7) used the look locker sequence and White et al. (10) looked at the myocardial and blood pool temporal nulling on TI scout sequence for diagnosing CA. Fontana et al. (11) used magnitude-IR sequence in majority of their patients (57%) to diagnose amyloidosis. The investigators used myocardial T1 values and extracellular volume using ShMOLLI sequence (pre-contrast and 15, 45, and 80 min following 0.0011 mmol/kg/min Gd injection) to offset errors in myocardial nulling prior to adoption of the PSIR sequence.
LGE and mortality
Across the 7 studies, there were 425 patients with 149 deaths during a mean follow-up period of 25 months. On pooled analysis, presence of LGE was associated with an OR of 4.96 (95% confidence interval [CI]: 1.90 to 12.93, 42% vs. 15%; p = 0.001) for all-cause mortality (Figure 2). There was no significant heterogeneity (p = 0.11) and, the I2 statistic indicated 43%. Much of the heterogeneity was in the same direction with 6 of 7 studies having ORs >1 with varying magnitudes. A sensitivity analysis indicated that the results were not greatly impacted by any 1 study; the ORs were consistent and the lower bound of the overall 95% CI remained >1 when data were reanalyzed 7 times excluding 1 study at a time. Forest plots were examined and Peters’ test (p = 0.714) and Egger’s test (p = 0.508) did not indicate significant publication bias. Follow-up ranged from 12 to 32 months among the 7 studies, which in total provided roughly 880 person-years of follow-up. For the LGE-negative group, the pooled death rate was 0.07 (95% CI: 0.03 to 0.19) events per year and for the LGE-positive group, the rate was 0.25 (95% CI: 0.16 to 0.39) per year (p = 0.001) (Figure 3). For 7 studies, the proportion of cardiac biopsy ranged from 3% to 68%. There was no apparent relationship between study-specific ORs relating LGE status to death and the proportion of patients with cardiac biopsy (Figure 4).
Type of CA and mortality
Three studies were composed solely of patients with AL amyloidosis (7–9) and one study included predominantly AL with only 2 patients with ATTR (10). These studies could not be included in an analysis investigating the association of amyloid type on death. Pooling data across the 3 remaining studies (Figure 5), there were 57 deaths in 148 patients with AL amyloid (39%) and 35 deaths in 144 patients with ATTR (24%). Patients with AL amyloid were more likely to die compared with those with ATTR, but this finding did not reach statistical significance (pooled OR: 2.13; 95% CI: 0.36 to 12.45; p = 0.40). There was substantial heterogeneity among these studies (p = 0.05); the I2 was 67%. With the information provided within studies, it was possible to investigate the effect of LGE status within AL amyloidosis across 4 studies (7–9,11) (Figure 6). Pooling data indicated that the results were consistent with the original analysis of increased mortality in LGE positive patients (pooled OR: 8.87; 95% CI: 2.08 to 37.81; p = 0.003). The heterogeneity among studies was not significant (p = 0.16); the I2 statistic was 41%.
CMR has developed as an imaging modality of choice for cardiac assessment including evaluation of left ventricular structure, function, perfusion, and tissue characteristics, including the presence or absence of LGE (18). The prognostic value of LGE by CMR has been well-documented in patients with nonischemic cardiomyopathy and hypertrophic cardiomyopathy (18,19). Similar prognostication in systemic amyloidosis has previously been demonstrated to be variable in single-center studies but never been confirmed in larger patient populations (5–11). This systematic review and meta-analysis is the first large-scale analysis to support the role of LGE-CMR in identifying patients with systemic amyloidosis with known or suspected CA to be at increased risk for all-cause mortality. The study suggests that LGE-CMR acts as an independent prognostic marker for outcomes in these patients irrespective of histopathological diagnosis of amyloidosis on cardiac biopsy. Recent studies have supported the incremental prognostic value of presence of LGE (11). LGE-CMR thus may serve the purpose of risk stratification of these patients for aggressive medical management possibly leading to a decrease in all-cause mortality. This holds more significance because LGE-CMR may detect early cardiac abnormalities in patients with amyloidosis with normal left ventricular thickness (4). Prior published studies have raised the question of whether a positive LGE-CMR result can obviate the need for endomyocardial biopsy for the diagnosis of cardiac involvement in patients with known amyloidosis. CMR would appear a reasonable alternative to endomyocardial biopsy in patients with tissue diagnosis from an alternate site and thought to be at high risk for invasive investigation, particularly where classic CMR findings of cardiac amyloid are identified (20,21). Our analysis suggests LGE as marker of increased mortality irrespective of cardiac biopsy results. Although LGE is the gold standard CMR technique in evaluation of cardiac amyloidosis, other newer techniques such as T1-mapping, intramyocardial T1 difference, look locker, TI scout, and equilibrium contrast-enhanced MR was successfully used to characterize diffuse myocardial fibrosis associated with amyloidosis. The role of these CMR techniques as a prognostic marker, similar to LGE, appears promising but needs further study and validation (22). Another interesting variable that probably affects prognosis is the underlying etiology for systemic amyloidosis. The study by Dungu et al. (23) showed that LGE-CMR patterns tend to differ in AL and ATTR. Ninety percent of ATTR patients demonstrate transmural LGE compared with 37% of AL patients. The study found survival was significantly better in cardiac ATTR amyloidosis compared with AL type. A recent large-scale prospective study by Fontana et al. (11) supported higher prevalence of transmural LGE in ATTR amyloidosis. The study found subendocardial LGE was more prevalent in AL (39% in AL vs. 24% in ATTR; p < 0.05) and transmural LGE more prevalent in ATTR amyloidosis (27% in AL vs. 63%; p < 0.0001). It is noteworthy that AL amyloidosis with transmural LGE did worse than ATTR amyloidosis. The chances of survival at 24 months were 61% with transmural LGE (45% in AL 65% in ATTR). Transmural LGE remained independent risk factor after adjusting for NT-pro BNP, ejection fraction and other variables. The study concluded a continuum of cardiac involvement in systemic AL and ATTR amyloidosis with transmural LGE representing advanced cardiac amyloidosis. There was no significant difference in mortality in AL versus ATTR amyloidosis with no LGE or subendocardial LGE. The median survival in patients with transmural LGE was 17 months in AL and 38 months in ATTR. Considering the above results, the degree of LGE enhancement (or “transmurality”) likely identifies a spectrum of disease progression and as such LGE-CMR appears an attractive noninvasive modality for early recognition and consequent institution of therapies that may favorably impact long-term outcomes. In our meta-analysis the incidence of death was higher for AL patients compared to ATTR amyloidosis (39% vs. 24%) albeit not statistically significant on account of fewer studies (3 of 7) exploring the difference in outcomes between the 2 subtypes. Another subgroup analysis within the AL amyloidosis population however confirmed that the LGE-positive population had worse outcomes. Different studies also used different techniques for detecting LGE, which itself could have prognostic implications. As mentioned previously, LGE is best evaluated using the PSIR sequence; however, this sequence was not universally available in all centers. Several studies used surrogate CMR techniques to offset the lack of PSIR. For example, Maceira et al. (6), who reported that LGE by itself was not a significant predictor of mortality, used intramyocardial T1 difference between subepicardium and subendocardium for diagnosis of cardiac amyloidosis. They found this technique was 85% accurate in predicting mortality in patients with systemic AL amyloidosis. However, their study had limitations. Apart from nonavailability of PSIR sequence, only 2 of 29 patients in their study had cardiac biopsy-proven amyloid. These factors may explain why LGE could not predict mortality in their study. Contrary to Maceira et al. (6), Migrino et al. (8) and Austin et al. (5) showed poor long-term survival in AL amyloidosis. These studies used PSIR sequence for the diagnosis of cardiac amyloidosis. Although the number of subjects in these studies was not large, the majority of patients in the study by Austin et al. (5) had biopsy-proven cardiac amyloidosis, giving additional credibility to their results. Mekinian et al. (7) and White et al. (10) also found increased risk of mortality in patients suspected of cardiac amyloidosis. However, both studies used a different approach to diagnose cardiac amyloid involvement. Mekinian et al. (7) defined the CMR-positive group based on lack of normal myocardial nulling on look locker sequence with inversion time >300 ms whereas White et al. (10) used a rapid visual assessment of myocardial and blood pool temporal nulling on TI scout (analogous to the look locker) sequence. The look locker and TI scout techniques are also based on gadolinium enhancement but instead of demonstrating delayed enhancement, they study gadolinium kinetics in the myocardium and blood pool. These techniques are most useful when the PSIR sequence for LGE is not available. In study by Fontana et al. (11), for the majority of the patients, PSIR sequence was not obtained. In the early part of enrollment, magnitude inversion recovery was used to diagnose cardiac amyloidosis. PSIR was obtained in only 43% of the subjects. The investigators used myocardial T1 values and extracellular volume using ShMOLLI sequence to offset errors in myocardial nulling prior to adoption of the PSIR sequence. The authors later concluded that the PSIR technique provides incremental prognostic information even after adjusting for known prognostic factors in patients with systemic AL and ATTR amyloidosis. In summary, the presence of LGE is associated with increased mortality in patients with systemic amyloidosis and known or suspected CA. It is pertinent to know that there is significant variation in outcomes based on pattern of LGE, type of MRI sequence used and underlying etiology of systemic amyloidosis. A multitude of factors including type, stage, comorbid conditions, candidacy for chemotherapy, and presence or absence of heart failure is likely to dictate overall outcomes.
The limitations of our study are predominantly related to those inherent to the included studies, which were all observational and predominantly single center. Although no obvious publication bias was apparent, the small number of studies limits a thorough assessment of publication bias. Second, most of the studies did not evaluate other relevant clinical endpoints such as sudden cardiac death, heart failure, and cardiovascular mortality in relation with presence of LGE-CMR though higher overall mortality was found in patients with underlying heart failure. There were limitations in the reported data provided in these observational studies in terms of patient demographics, relevant clinical findings, staging of the disease, and medical management including use of chemotherapeutic drugs and autologous bone marrow transplantation, and therefore, the meta-analysis and the pooled estimates could not be adjusted for potential confounders. Finally, patient-level data were not available precluding subgroup analysis and meta-regression.
The results of this systematic review and meta-analysis support the prognostic role of LGE-CMR in risk stratification of patients with CA.
COMPETENCY IN MEDICAL KNOWLEDGE: CMR has emerged as the noninvasive imaging modality of choice for the diagnosis of cardiac amyloidosis. In these patients, the presence of LGE provides incremental prognostic utility with an increase in all-cause mortality seen over time. These patients have an increased risk of adverse events irrespective of cardiac biopsy results.
TRANSLATIONAL OUTLOOK: These findings need to be validated by other long-term prospective studies involving larger patient numbers. Additional CMR findings need to be tested for their prognostic capabilities in patients with cardiac amyloidosis. In these patients, the information provided by CMR can help guide future therapies to prevent adverse outcomes.
All authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- amyloid light-chain
- cardiac amyloidosis
- confidence interval
- cardiac magnetic resonance
- late gadolinium enhancement
- left ventricular ejection fraction
- magnetic resonance imaging
- New York Heart Association
- odds ratio
- phase-sensitive inversion recovery
- Received November 30, 2015.
- Revision received January 11, 2016.
- Accepted January 14, 2016.
- American College of Cardiology Foundation
- Kyle R.A.
- Syed I.S.,
- Glockner J.F.,
- Feng D.,
- et al.
- Austin B.A.,
- Tang W.H.,
- Rodriguez E.R.,
- et al.
- White J.A.,
- Kim H.W.,
- Shah D.,
- et al.
- Fontana M.,
- Pica S.,
- Reant P.,
- et al.
- Kuruvilla S.,
- Adenaw N.,
- Katwal A.B.,
- Lipinski M.J.,
- Kramer C.M.,
- Salerno M.
- Selvanayagam J.B.,
- Leong D.P.
- Saeed M.,
- Van T.A.,
- Krug R.,
- Hetts S.W.,
- Wilson M.W.
- Dungu J.N.,
- Valencia O.,
- Pinney J.H.,
- et al.