Author + information
- Received October 4, 2018
- Revision received October 25, 2018
- Accepted October 25, 2018
- Published online January 28, 2019.
- Giuseppe D. Sanna, MD, PhDa,∗ (, )
- Giuseppe Nusdeo, MDa,
- Maria Rita Piras, MDb,
- Antonietta Forteleoni, MDa,
- Maria Rita Murru, BSc,
- Pier Sergio Saba, MD, PhDa,
- Simone Dore, PhDd,
- Giovanni Sotgiu, MD, PhDd,
- Guido Parodi, MD, PhDe and
- Antonello Ganau, MDe
- aClinical and Interventional Cardiology, Sassari University Hospital, Sassari, Italy
- bUnità di valutazione Alzheimer, Azienda Ospedaliero Universitaria Sassari, Ospedale San Camillo, Sassari, Italy
- cLaboratorio Centro Sclerosi Multipla, Università di Cagliari, Cagliari, Italy
- dClinical Epidemiology and Medical Statistics Unit, Department of Biomedical Sciences, University of Sassari, Sassari, Italy
- eUniversity of Sassari, Sassari, Italy
- ↵∗Address for correspondence:
Dr. Giuseppe Damiano Sanna, Clinical and Interventional Cardiology, Sassari University Hospital, Sassari (Italy), Via Enrico De Nicola, 07100 Sassari, Italy.
Objectives This case control study sought to assess the presence and characteristics of cardiac abnormalities in patients with Alzheimer disease (AD).
Background Protein misfolding is involved in the pathophysiology of neurodegenerative disorders such as AD. Recently, amyloid-beta (Aβ) aggregates were identified within the cardiomyocytes and interstitium of patients with AD, suggesting that Aβ oligomers may reach and damage the heart.
Methods The authors studied 32 patients with AD and 34 controls matched by age and sex, all of whom were free from cardiac or systemic diseases. A clinical evaluation, an electrocardiogram, and an echocardiogram were performed in all subjects. Furthermore, patients with AD underwent genetic analyses (of the PSEN1, PSEN2, APP, and APOE genes).
Results Compared to the control group, patients with AD had a higher prevalence of low-voltage electrocardiographic QRS complexes (28% vs. 3%, respectively; p = 0.004), a lower voltage/mass ratio (p = 0.05), a greater echocardiographic interventricular septum (10.1 ± 1.3 mm vs. 9.3 ± 1.1 mm, respectively; p = 0.01), a greater maximum wall thickness (10.8 ± 1.7 mm vs. 9.3 ± 1.1 mm, respectively; p = 0.0001), and a 2-fold higher prevalence of diastolic dysfunction (70% vs. 35%, respectively; p = 0.007). Symptoms and signs of heart failure were absent in all patients with AD.
Conclusions This study shows that electrocardiographic and echocardiographic abnormalities, including diastolic dysfunction, are present in patients with AD and that these studies reproduce the pattern of cardiac amyloidosis. These findings suggest that, in AD, there may be subclinical cardiac involvement likely associated with Aβ amyloid deposition. The clinical relevance of these cardiac abnormalities should be evaluated in larger prospective studies.
In a worldwide population that is aging, there is a parallel increase in chronic and degenerative diseases. In addition to cardiovascular diseases, major neurocognitive disorders (e.g., dementia) significantly contribute to the health care-associated burden (1). Alzheimer Disease (AD) is the most common form of dementia (2).
Protein misfolding plays a central role in the pathogenesis of neurodegenerative diseases, particularly in AD (3). In the past few years, it has been suggested that AD shares this mechanism with cardiac diseases (4–6). Mutations in the presenilin 1 (PSEN1) and PSEN2 genes, associated with familial forms of AD, have been found in patients with dilated cardiomyopathy (4,5,7). Even the ε4 variant of the apolipoprotein E (APOE) gene, a risk factor for late-onset AD (8), has been found in patients with dilated cardiomyopathy (9). More recently, Aβ amyloid aggregates (Aβ40 and Aβ42) were identified within cardiomyocytes and interstitial samples from dysfunctional myocardium in 4 patients with AD (10). Although these data suggest that protein misfolding may affect both the brain and the heart in patients with AD, the clinical evidence and characteristics of cardiac involvement in AD are still lacking. The present study investigated the presence and characteristics of structural and functional cardiac abnormalities in patients with AD.
The present study is a case control study with a convenience sample of patients who were consecutively enrolled in order to assess the possible association between AD and cardiac abnormalities.
Ethical approval and consent
The study protocol was approved by the local Research Ethics Committee. Written informed consent was obtained from the patients or from their legal guardians.
We analyzed the clinical records of 108 patients with AD referred to the Alzheimer Evaluation Unit of the University Hospital of Sassari, Sassari, Italy, who had a disease duration of at least 3 years. The clinical condition was diagnosed following criteria of the Diagnostic and Statistical Manual (DSM) of Mental Disorders IV-TR (11). Furthermore, in a significant percentage of cases, single-photon emission computed tomography (SPECT) was available and was used as an adjunct test to confirm the clinical diagnosis and to differentiate AD from other dementia types on the basis of differences in perfusion patterns. Of 108 AD screened individuals, 50 had at least 1 of the following exclusion criteria introduced to avoid possible confounding effects of systemic or cardiovascular comorbidities: malignant tumors (n = 4), systemic inflammatory diseases (n = 12), amyloidosis (n = 3), known or clinically suspected coronary artery disease (n = 11), moderate to severe valvular disease (n = 8), grade II to III arterial hypertension (blood pressure >160/100 mm Hg) (n = 10), pericardial disease (n = 1), and congenital heart disease (n = 1). Uncomplicated grade I arterial hypertension (blood pressure <160/100 mm Hg) was not considered an exclusion criterion because high blood pressure is common in elderly subjects. An additional 26 patients could not be enrolled due to a denial of informed consent (12 cases), a refusal of transport to the echocardiography laboratory due to severe disability (6 cases), or a lack of compliance with echocardiographic examination (8 cases).
The final study group included 32 subjects who had AD (in 16 subjects a positive SPECT study was available as support for clinical diagnosis) and were free of cardiovascular or systemic disease. The control group included 34 subjects matched by age and sex who were waiting for minor noncardiac surgery. These subjects were consecutively enrolled by using the same exclusion criteria adopted for the AD group.
Clinical evaluation and electrocardiography
Patient medical histories were recorded, and complete physical and cardiovascular examinations were performed in all study subjects. Close relatives of AD patients were consulted to integrate medical history and records. Patients underwent a standard 12-lead electrocardiography (ECG) study and examinations of PR and QTc time intervals. QRS duration and voltages were measured. Cardiac hypertrophy was defined according to Sokolow-Lyon or Cornell voltage criteria or both. QRS voltage was considered low when QRS amplitude was ≤0.5 mV in limb leads or ≤1 mV in precordial leads. The total QRS score (sum of QRS voltages in the limb and precordial leads) was also measured. Voltages were measured in millimeters after calibration (1 mm = 0.1 mV). Voltage/mass ratios were calculated as Sokolow-Lyon indexes, (mm)/left ventricle (LV) cross-sectional area index (12), and as total QRS scores (mm)/LV mass index (g/m2.7).
ECG measurements were repeated by 2 experienced and blinded operators (G.D.S. and G.N.) in order to limit potential sources of bias. Average measurements were considered.
Two-dimensional echocardiography was performed using a Vivid 7 machine (GE Medical Systems, Milwaukee, Wisconsin). Chamber quantification was performed according to the protocols of the European Association of Echocardiography and American Society of Echocardiography (13). LV diameters were measured from a parasternal long-axis view, using 2-dimensional (2D)-guided M-mode images. Interventricular septum and posterior wall thicknesses were measured using 2D-guided M-mode images, either from the parasternal long-axis or the short-axis view; the greatest measurement was defined as maximum wall thickness. LV hypertrophy was assessed according to sex-specific values of LV mass indexed by height (m2.7) ≥48 g/m2.7 for men and ≥44 g/m2.7 for women (13).
LV remodeling was defined as a normal LV mass index with relative wall thickness of ≥0.44 mm for both sexes (14). Left atrial volume was measured using the area-length technique (13). Severity of both the mitral and aortic regurgitation was assessed by evaluating color flow imaging and “vena contracta” width (15).
Diastolic function was assessed by using pulsed wave Doppler across the mitral valve to measure the E/A ratio and by tissue Doppler velocity imaging at the level of the mitral valve annulus (both septal and lateral) to obtain the average E/e′ ratio.
All echocardiographic measurements were performed offline by 2 experienced and blinded operators (G.D.S. and A.G.) using dedicated software (OsiriX, Pixmeo Co., Geneva, Switzerland), and average measurements were considered. Both of the operators independently analyzed both cases and controls (all images).
Blood samples for genetic testing were analyzed at the research laboratory of the Multiple Sclerosis Center, University of Cagliari, Cagliari, Italy. DNA was extracted by using the salting-out method, and polymerase chain reaction was carried out using specific primers. Amplified genes and fragments were extracted from the APOE (19q13.2), APP (21q21), PSEN1 (14q24.3), and PSEN2 (1q31-q42) genes. The Alzheimer Disease and Frontotemporal Dementia Mutation Database (16) was used to verify sequence variants.
A structured follow-up examination by telephone interview was planned for all subjects included in the study, within 3 years from enrollment.
Statistical analyses were performed using Stata13 software (StataCorp, College Station, Texas). The Student’s t-test and Mann-Whitney U test or the chi-square test was used to compare quantitative and qualitative data, respectively. The values of p < 0.05 were considered statistically significant.
Relationships between age and both the maximum wall thickness and the E/A ratio were analyzed, and Pearson's or Spearman's correlation coefficients were obtained. Forward stepwise multivariate analysis was performed in patients with AD to assess the independent determinants of maximum wall thickness by entering into the model a set of 7 variables with potential pathophysiological relevance (AD, age, arterial hypertension, AD duration, body mass index, diabetes, and sex).
Inter-reader variability of ECGs and echocardiographic measurements was tested by using Bland-Altman analysis, which showed good interobserver agreement. The alpha value was 0.05 for all analyses.
Anthropometric and clinical characteristics of the study patients are summarized in Table 1. Age, sex, and prevalence of cardiovascular risk factors did not significantly differ between patients with AD and controls. The 2 groups were also taking similar cardiac medications. A total of 14 of 32 patients with AD were familiar with their cognitive disorders (44%).
ECG characteristics are reported in Table 2. No statistically significant differences between AD patients and controls were detected regarding rhythm abnormalities or lengths of PR or QTc intervals, conduction disorders, pathologic Q waves, repolarization abnormalities, or LV hypertrophy. The AD group had a much higher prevalence of low-voltage QRS complexes (p = 0.004) and lower total QRS scores (p = 0.05) than the control group (Table 2, Online Figure 1). Both chronic obstructive pulmonary disease and obesity, both of which are potential causes of low-voltage QRS complexes, were distributed equally in each of the 2 groups.
Echocardiographic characteristics are reported in Table 3. LV end-diastolic diameter, relative wall thickness, mass index, and ejection fraction in the patients with AD were not statistically different from those in the control group. However, patients with AD showed greater maximum wall thickness (10.8 ± 1.7 mm in patients with AD vs. 9.3 ± 1.1 mm in controls; p = 0.0001) and greater interventricular septum (p = 0.01) than those in controls (Online Figure 2). Because LV wall thickness can be influenced by several factors (e.g., AD, age, sex, arterial hypertension, obesity, diabetes mellitus), multivariate regression analysis was carried out to assess the role of each variable as an independent predictor of LV maximum wall thickness (Table 4). AD, age, and arterial hypertension were independent predictors of maximum wall thickness and cumulatively explained 49% of its global variability; AD duration, body mass index, diabetes, and sex did not enter the model. A significant and direct relationship between age and maximum wall thickness was found in the AD group (r = 0.44; p = 0.01) but not in the control group (Online Figure 3).
The 2 groups had normal ejection fractions, whereas the AD group showed significantly lower E/A ratios (p = 0.02) and a trend toward a higher E/e′ ratio than the control group (Table 3). Consistently, the prevalence of diastolic dysfunction was twice as high in patients with AD as it was in the control group (70% vs. 35%, respectively; p = 0.007) (Online Figure 4).
A significant, direct relationship between the E/e′ ratio and age was evident in AD patients (r: 0.62; p = 0.0008) but not in controls (Online Figure 5).
Combining QRS voltages with LV mass showed AD patients had lower total QRS scores/LV mass index ratios (p = 0.05) than controls (Table 2).
A single missense mutation in the PSEN2 gene was observed in a female patient with familial late-onset AD. Two subjects were carriers of the Glu318Gly polymorphism of the PSEN1 gene. Two subjects were homozygous for the ε4 allele of the APOE gene, whereas 10 subjects were heterozygous for that same allele. One of those 10 subjects was also a carrier of the Glu318Gly polymorphism of the PSEN1 gene. In 5 subjects, the PSEN1 gene polymorphism was associated with the PSEN2 gene polymorphism (Online Table 1).
Median (interquartile range [IQR]) follow-up was performed at 39 (IQR: 39 to 43) months in the AD group and at 47 (IQR: 43 to 56.5) months in the control group. Follow-up data were available in 31 of 32 patients with AD and 28 of 34 control subjects. In the AD group, there were 9 noncardiovascular deaths and 1 death from heart failure. In the control group, there were 2 noncardiovascular deaths and 2 nonfatal cardiovascular events (Online Table 2).
Cardiovascular disease is the second leading cause of death after pneumonia in patients with AD (17). Cardiovascular risk factors have been associated with the risk of vascular dementia (18), but their relevance is less clear in AD. Recently, an autopsy study of 4 AD patients demonstrated the presence of Aβ amyloid also in the myocardium, raising a new pathogenic hypothesis of AD as a multiorgan disease caused by protein misfolding (10). Some investigators have reported a correlation between dilated cardiomyopathy and some AD-related genes (4,5).
Data from the present study do not support the hypothesis of an association between AD and dilated cardiomyopathy (Table 3). The present authors found a significant increase in LV wall thickness (Table 3, Online Figure 1) and a trend toward concentric remodeling in AD patients compared to the control group. AD patients were older, and it is known that aging induces concentric LV remodeling (19). Furthermore, hypertension was largely represented in the present sample, and hypertension affects LV geometry through increases in absolute and relative wall thicknesses (14). However, the 2 groups did not differ significantly by age, history of hypertension, and blood pressure levels. In multivariate regression analysis, AD was the major independent predictor of maximum wall thickness, followed by age and hypertension. These data, taken together with the significant direct relationship between wall thickness and age (Online Figures 4 and 5), suggest an interaction among AD, aging, and hypertension in contributing to the ventricular remodeling trend.
In contrast to the increase in LV wall thickness, the ECG in the AD group revealed lower peripheral QRS voltages (Online Figure 1) and total QRS scores (Table 2). Moreover, the total QRS score/LV mass index ratio and the Sokolow-Lyon voltage/LV cross-sectional area ratio were also lower in the AD group than in the control group (p = 0.05 and 0.06, respectively) (Table 2). These findings of low ECG QRS voltages and voltage/mass ratio indices have never been previously reported in AD. Another significant difference was the higher prevalence of diastolic dysfunction in AD patients than in control patients (70% vs. 35%, respectively; p = 0.007) (Table 3, Online Figure 4).
The present data partially agree with recent observations of Troncone et al. (10). Using a smaller data set, their study found increased LV wall thickness only in older patients with AD rather than in the control groups. Troncone et al. (10) also found a reduction in the E/A ratio in younger AD subjects that progressively overlapped with that in the controls at advanced ages, suggesting independence of diastolic dysfunction from wall thickening. In the study by Troncone et al. (10), the more accurate index of diastolic function E/e′ ratio was not available (10).
In AD patients in the present study, the LV wall thickness increased progressively with age. Unlike observations in the study by Troncone et al. (10), the E/A ratio in the present study was reduced in the AD group, regardless of age, whereas the E/e′ ratio increased significantly with age (Online Figure 5). Present data suggest a synergistic action between AD and aging in inducing both wall thickening and diastolic dysfunction.
In 4 patients with AD studied autoptically, Troncone et al. (10) identified Aβ amyloid aggregates in the brain and also in cardiomyocytes and cardiac interstitium. Accordingly, Troncone et al. (10) proposed AD is a multiorgan disease due to protein misfolding, with circulating Aβ peptides representing the unifying mechanism of disease involving brain and heart.
According to that pathophysiological framework, findings from the present study point to a mild cardiac involvement in patients with AD, similar to the early phase of cardiac amyloidosis. The low QRS voltages and the disproportion between LV wall thickness and QRS voltages (the voltage/mass ratio) present in this AD group (Table 2) are well recognized hallmarks of cardiac amyloidosis (20). The thicker LV walls and higher prevalence of diastolic dysfunction also contributed to support this interpretation.
In the present study, arterial hypertension was an independent predictor of maximum wall thickness. Recent data obtained from an animal model of transverse aortic coarctation showed that high central blood pressure may induce production and deposition of Aβ amyloid substance, both in the brain and in the heart (21). This suggests a synergistic interaction between high blood pressure and AD in inducing the deposition of Aβ oligomers and myocardial abnormalities.
The independent association between age and wall thickening may reflect the longer duration of exposure of the heart to Aβ oligomers, possibly enhanced by the high blood pressure levels that are very common in advanced age.
Diabetes may also contribute to stimulating the aggregation of Aβ and tau proteins secondarily to the formation of advanced glycation end products (22,23). However, diabetes was similarly represented in the present AD and control groups.
In cardiac amyloidosis, cardiac walls are usually thicker and symmetrical, particularly in the wild-type transthyretin-related (ATTRwt) form (24), and ejection fraction is usually reduced. Moreover, the diastolic dysfunction observed in the present patients was less severe than in cardiac amyloidosis (25).
This suggests that the cardiac damage in patients with AD is mild or characterized by slow progression. In the latter case, death due to the neurodegenerative process and related complications could precede the cardiovascular events. The follow-up data from the present patients seem to confirm this hypothesis.
The relatively mild alterations observed in the present study must take into account the fact that the present group of patients was selected to be asymptomatic and free of cardiovascular diseases, to avoid the confounding effects of comorbidities. Therefore, patients with more severe cardiac impairment could have been excluded.
Polymorphisms of the PSEN1 and PSEN2 genes were found in 5 AD patients, and none of them was associated with dilated cardiomyopathy, as suggested by other reports (4,5) or peculiar cardiac patterns.
The possible link between AD and myocardial disease raises interesting research perspectives in treating the 2 coexisting diseases. New therapeutic strategies targeting Aβ oligomers may help to treat both groups of disorders. Immunotherapies with oligomer-specific antibodies have given promising results in the experimental model of AD, but they must be tested specifically in humans, alone or associated with other treatments (26).
Postmortem biopsy represents the gold standard technique for diagnosis of AD; however, presently, the diagnosis worldwide is clinically based on DSM criteria, and nuclear imaging (i.e., SPECT) may be useful. In order to increase accuracy, the present study enrolled AD patients who were being managed in a tertiary university referral center.
A limitation of the present study was the small sample size, due to the intrinsic difficulties in both obtaining informed consent from patients or tutors and performing echocardiograms in very poorly compliant patients. The high prevalence of comorbidities also contributed to limiting the sample size. Furthermore, larger studies are needed to define the whole spectrum of heart abnormalities in AD, including symptomatic patients and those with a longer duration of illness.
In the present study, echocardiographic scans were performed by 1 of 2 echocardiographic experts (i.e., not double-blinded), and this fact could have somehow introduced a bias.
Finally, echocardiography does not provide information about tissue composition. Speckle tracking echocardiography has been used in cases of cardiac amyloidosis to evaluate myocardial deformation (27), but it provides only functional information. Cardiac magnetic resonance imaging using the newer T1 or T2 extracellular volume mapping techniques can detect myocardial damage at an early stage (28,29) but cannot be used in noncompliant AD patients, as in the present cohort. It was not possible to perform cardiac magnetic resonance imaging in these patients due to the advanced stage of dementia requiring general anesthesia in order to complete the scan.
This study revealed a peculiar involvement of the heart in AD, an involvement which is characterized by a generally mild increase in interventricular septum thickness, diastolic dysfunction, and ECG abnormalities including reduced QRS voltages and voltage/mass ratios. These amyloidosis-like features are associated with slow progression over years and a limited number of cardiac events during follow-up. Because only asymptomatic patients were selected and free of known heart disease, the alterations observed can be attributed only to the underlying disease.
The present data are consistent with the recent evidence that Aβ aggregates propagate outside the brain and reach the heart (10), inducing AD-related cardiac amyloidosis (Figure 1). However, the clinical significance of these structural abnormalities and instrumental findings should be demonstrated, provided that overall outcome in our small cohort did not differ significantly between patients with AD and controls in terms of cardiovascular events. Clinicians should pay attention to cardiac abnormalities when evaluating patients with AD.
Larger studies may help clarify the full spectrum of myocardial involvement in AD and its real clinical relevance.
COMPETENCY IN MEDICAL KNOWLEDGE: Patients with AD often exhibit increased cardiac wall thickness and diastolic dysfunction, as assessed by echocardiography, low-QRS voltages on ECG, and low voltage/mass ratios compared to those in controls. This amyloid-like cardiac pattern reflects a peculiar involvement of the heart and suggests a shared disease mechanism (i.e., deposition of the same material resulting from protein misfolding) that affects both brain and heart in AD. Clinicians should pay attention to subtle cardiac abnormalities when evaluating patients with AD.
TRANSLATIONAL OUTLOOK: The presence of cardiomyopathy in AD should be evaluated in larger studies. A common pathophysiologic process underlying neurodegenerative and cardiac diseases may justify the development of common therapeutic strategies targeted at Aβ oligomers to treat both groups of disorders.
The authors thank Federica Decandia and Giuseppe Colletti for assistance in data collection, and Raffaele Murru for genetic analysis.
The authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- Alzheimer disease
- amyloid precursor protein
- apolipoprotein E
- left ventricle
- Received October 4, 2018.
- Revision received October 25, 2018.
- Accepted October 25, 2018.
- 2019 American College of Cardiology Foundation
- American Psychiatric Association
- Gianni D.,
- Li A.,
- Tesco G.,
- McKay K.M.,
- Moore J.,
- et al.
- Corder E.H.,
- Saunders A.M.,
- Strittmatter W.J.,
- et al.
- Troncone L.,
- Luciani M.,
- Coggins M.,
- et al.
- American Psychiatric Association
- Lang R.M.,
- Bierig M.,
- Devereux R.B.,
- et al.
- Ganau A.,
- Devereux R.B.,
- Roman M.J.,
- et al.
- Ganau A.,
- Saba P.S.,
- Roman M.J.,
- et al.
- Maurer M.S.,
- Elliott P.,
- Comenzo R.,
- Semigran M.,
- Rapezzi C.
- Carnevale D.,
- Lembo G.
- Rios J.A.,
- Cisternas P.,
- Arrese M.,
- et al.
- Quarta C.C.,
- Solomon S.D.,
- Uraizee I.,
- et al.
- Rapezzi C.,
- Merlini G.,
- Quarta C.C.,
- et al.
- Sengupta U.,
- Nilson A.S.,
- Kayed R.
- D'hooge J.,
- Heimdal A.,
- Jamal F.,
- et al.
- Moon J.C.,
- Messroghli D.R.,
- Kellman P.,
- et al.
- Fontana M.,
- Banypersad S.M.,
- Treibel T.A.,
- et al.