European Journal of Heart Failure (2017)

doi:10.1002/ejhf.899

Inotropic myocardial reserve deficiency is the

predominant feature of exercise
haemodynamics in cardiac amyloidosis
Tor Skibsted Clemmensen1*, Henning Mølgaard1, Jens Sörensen2, Hans Eiskjaer1,
Niels Frost Andersen3, Søren Mellemkjaer1, Mads Jønsson Andersen1,
Lars Poulsen Tolbod2, Hendrik J. Harms2, and Steen Hvitfeldt Poulsen1
1 Department of Cardiology; 2 Department of Nuclear Medicine and PET Centre; and 3 Department of Haematology, Aarhus University Hospital, Aarhus, Denmark

Received 28 December 2016; revised 20 April 2017; accepted 2 May 2017

Aims This study aimed to characterize invasive haemodynamics during exercise in subjects with cardiac amyloidosis (CA).
.....................................................................................................................................................................
Methods The study population numbered 44 subjects. Group A (CA-positive, n = 24) comprised wild-type transthyretin
and results patients (n = 10), familial transthyretin amyloidosis mutation carriers (ATTRm) with cardiac involvement (n = 5),
and light-chain amyloidosis patients with cardiac involvement (n = 9). Group B (CA-negative, n = 20) comprised four
healthy ATTRm subjects without cardiac involvement documented by 11 C-PIB positron emission tomography and
16 healthy controls. All subjects underwent a symptom-limited, semi-supine exercise test with expired gas analysis
and simultaneous right heart catheterization. CA patients had lower peak oxygen consumption [15 ± 6 mL/min/kg
bodyweight (bwt) vs. 33 ± 7 mL/min/kg bwt; P < 0.0001] than controls. Myocardial reserve during exercise was
significantly reduced in CA patients as reflected in a small increase in stroke volume index (SVI) and cardiac index (CI)
compared with controls [ΔSVI: 4 mL/m2 (range: −1 to 8) vs. 14 mL/m2 (range: 5–25); P < 0.0001; ΔCI: 2 ± 2 L/min vs.
7 ± 2 L/min; P < 0.0001]. During exercise, CA patients had significantly higher left and right ventricular filling pressures
than controls. Furthermore, CA patients had severely impaired pulmonary arterial compliance (PAC) compared with
controls [2.9 mL/mmHg (range: 2.1–4.5) vs. 7.5 mL/mmHg (range: 5.7–10.4); P < 0.0001].
.....................................................................................................................................................................
Conclusions Cardiac amyloid deposits are associated with severely reduced inotropic myocardial reserve and increased left and
right ventricular filling pressures during exercise. Furthermore, CA subjects have severely reduced PAC, which may
contribute to right heart failure and reduced exercise capacity.
..........................................................................................................
Keywords Cardiac amyloidosis • Haemodynamics • Heart failure • Positron emission tomography •
Exercise

Introduction was strongly associated with long-term prognosis.4 – 6 Likewise,

........................

previous studies have demonstrated severely reduced exercise

Cardiac amyloidosis (CA) is characterized by extracellular deposi-
tion of misfolded proteins that leads to increased biventricular wall
thickness and myocardial stiffness, and eventually results in heart
failure and death.1 – 3 Invasive assessment of myocardial filling using
right heart catheterization (RHC) during rest has revealed signifi-
cantly elevated biventricular filling pressures in CA patients, which
capacity as measured by peak oxygen consumption (VO2 ) and
exercise duration in CA patients.7 However, the link between
impaired exercise capacity and haemodynamic alterations during
functional stress has not been studied. Most patients with diastolic
dysfunction will have a normal or mildly elevated left ventricular
(LV) filling pressure at rest. In these patients, symptoms will develop

*Corresponding author. Department of Cardiology, Aarhus University Hospital Skejby, Palle Juul-Jensens Boulevard 99, 8200 Aarhus N, Denmark. Tel: +45 78 45 22 51,
Fax: +45 78 45 22 60, Email: torclemm@rm.dk

© 2017 The Authors

European Journal of Heart Failure © 2017 European Society of Cardiology
2 T.S. Clemmensen et al.

with physical activity only because of rising filling pressures, which (SVI). Pulmonary arterial compliance (PAC) was calculated as:

........................................................................................................................................................................
are necessary to maintain adequate stroke volume (SV). Therefore, SV/(sPAP − dPAP). The pulmonary vascular resistance index (PVRI)
the evaluation of central haemodynamics at rest may fall short of was calculated as: PVRI = 80 × (mPAP − mPCWP)/CI. The systemic
explaining the haemodynamic mechanisms leading to symptomatic vascular resistance index (SVRI) was calculated as: SVRI = 80 × [mean
arterial pressure (MAP) − mRAP]/CI. Arteriovenous oxygen (O2 )
CA. Thus, the present study aimed to determine invasive sys-
content difference was measured as the difference between pulse
tolic and diastolic performance at rest and during exercise in CA
oxymetry measurements and directly measured PA O2 content at
patients compared with healthy individuals, and to study the rela- rest and during peak exercise. Right ventricular stroke work index
tionship between invasive exercise haemodynamics and exercise (RVSWI) was calculated as: RVSWI = SVI × 0.0136 × (mPAP − mRAP).
capacity. Similarly, LV stroke work index (LVSWI) was calculated as
LVSWI = SVI × 0.0136 × (MAP − mPCWP). LV preload adjusted stroke
work (LV-PASW) was calculated as: LV-PASW = (SV × MAP)/two-dim-
Methods ensional (2-D) end-diastolic volume (EDV) (from echocardiography).
Based on previous studies in healthy individuals, a resting mPAP of
Patients >20 mmHg and PCWP of >15 mmHg were considered to be abnormal.
At peak exercise, an mPAP of ≥30 mmHg and PCWP of ≥25 mmHg
The study was carried out at the Department of Cardiology, Aarhus
were considered to be abnormally increased.11 – 13
University Hospital, Denmark from September 2015 to November
2016. A total of 44 subjects were enrolled. Of these, 24 subjects
(CA-positive; Group A) had confirmed CA and 20 subjects served
as controls (CA-negative; Group B). Group A comprised wild-type Exercise protocol
transthyretin (ATTRwt) patients (n = 10), familial transthyretin amy- All patients performed a multistage symptom-limited, semi-supine
loidosis mutation carriers (ATTRm) of the Danish ATTRm mutation bicycle exercise test using the Echo Cardiac Stress Table (Lode BV,
(Leu111Met) with cardiac involvement (n = 5), and light-chain amyloi- Groningen, the Netherlands). The workload started at 0 W and
dosis (AL) patients with cardiac involvement (n = 9). Group B consisted was increased by 25 W every 3 min in patients of New York Heart
of healthy controls (n = 16) and healthy ATTRm carriers without car- Association (NYHA) class II or lower status, and by 10 W in patients
diac involvement as documented by 11 C-Pittsburgh Compound B (PIB) of NYHA class III or higher status. Patients were encouraged to
positron emission tomography (PET) (n = 4).8 – 10 The healthy con- maintain a fixed pedalling speed of 60 rounds/min and to exercise until
trols were recruited using posters displayed in the hospital facility and exhaustion (Borg >18).
consisted of hospital employees and patients’ or employees’ relatives. Peak oxygen consumption was measured using breath-by-breath
These healthy controls and healthy mutation carriers received no medi- analysis of expired gas at rest and at each step during exercise.
cation, had no cardiopulmonary symptoms and had normal electrocar- The ratio between ΔCO by thermodilution (peak CO − rest CO)
diograms. Patients were recruited from the department’s outpatient and ΔVO2 (peak VO2 − rest VO2 ) was used as a marker of cardiac
clinic. Patients with significant valve disease were excluded. Four of the response to metabolic needs.14 Oxygen pulse (OP) was calculated as
ATTRm patients with cardiac involvement had undergone liver trans- OP = VO2 /HR.
plant in order to delay amyloid progression. The fifth ATTRm patient
had not undergone endomyocardial biopsy, but was considered to have
CA because the findings of 11 C-PIB PET were positive. Echocardiography
All participants were aged ≥18 years and provided written informed
Echocardiography was performed at rest using a commercially available
consent according to the principles of the Helsinki Declaration before
ultrasound system (Vivid E95; GE Vingmed Ultrasound AS, Horten,
their inclusion in the study. The study was approved by the local science
Norway) with a 3.5-MHz phased array transducer (M5S) for 2-D eval-
ethics committee of Central Denmark Region.
uation and a 4 V-D transducer for three-dimensional (3-D) evaluation.
Assessed material included 2-D cine loops and tissue Doppler images
from all three apical views of the left ventricle, along with pulsed wave
Invasive haemodynamic measurements Doppler of the mitral inflow and the LV outflow tract. The LV fill-
Right heart catheterization was performed in all patients and con- ing pattern was characterized as normal, intermediate or of diastolic
trol subjects. A standard 7.5-F triple-lumen Swan–Ganz thermistor dysfunction using the American Society of Echocardiography (ASE)
and balloon-tipped catheter (Edwards Lifesciences, Irvine, CA, USA) guidelines.15 Long-axis displacement was estimated by tissue tracking
were used. The catheter was introduced into the right jugular vein using colour-coded tissue velocity images. This method displays the
under ultrasound guidance using the Seldinger technique, and advanced systolic tissue velocity integral, which equals the long-axis distance of
using pressure waveform and fluoroscopy into the pulmonary artery motion.16,17 Peak systolic myocardial displacement was determined as
(PA). Pulmonary capillary wedge pressure (PCWP), mean right atrial an average of basal septal, lateral, anterior and posterior tissue dis-
pressure (mRAP), LV transmural pressure (LVTMP = PCWP − mRAP), placement prior to aortic valve closure.
systolic and diastolic PA pressure (sPAP, dPAP), mean PA pres- Data on 3-D mass were obtained by sampling six heart beats during
sure (mPAP), transpulmonary gradient (TPG = mPAP − PCWP), car- breath-hold with a target frame rate of >25 frames/s. Both diastolic and
diac output (CO) and blood pressure were measured at rest, at each systolic LV mass were measured. An average of the two measurements
exercise level until exhaustion, and at 5 min post-exercise. was used as a marker of amyloid burden.
Cardiac output was measured using thermodilution as the aver- At peak exercise, 2-D images of volume changes were obtained
age of two measurements not differing by more than 10%. SV using the Simpson biplane method of discs. Data were analysed offline
was calculated as CO/heart rate (HR). CO and SV were indexed using dedicated software (EchoPAC PC SW-Only, 201; GE Healthcare,
to body surface area (BSA) as cardiac index (CI) and SV index Milwaukee, WI, USA).

© 2017 The Authors

European Journal of Heart Failure © 2017 European Society of Cardiology
Exercise haemodynamics in cardiac amyloidosis 3

Statistical methods Furthermore, CA patients had a severely reduced inotropic

........................................................................................................................................................................
myocardial reserve. As Figure 1 and Table 2 show, only small
Normally distributed data are presented as the mean ± standard devi-
ation (SD); non-normally distributed data are presented as the median exercise-induced increases in LVSWI and LV-PASW were seen in
and interquartile range (IQR). Categorical data are presented as abso- CA patients. The poor LVSWI and LV-PASW reserve was mainly
lute values with percentages. Histograms and Q–Q plots were used attributable to only a small increase in SV during exercise (Figure 2).
to check continuous values for normality. Between-group differences Thus, median ΔSVI values were 4 mL/m2 (IQR: −1 to 8 mL/m2 ) vs.
were assessed using the t-test for normally distributed data, the 14 mL/m2 (IQR: 5–25 mL/m2 ) and mean ± SD ΔCI was 2 ± 2 L/min
Mann–Whitney U-test for non-normally distributed data, and the 𝜒 2 vs. 7 ± 2 L/min (P < 0.0001) (P-values adjusted for baseline levels,
test for dichotomized data. Age is an important confounder in relation both <0.0001). In CA subjects, there was no relation between
to haemodynamics.18 Therefore, P-values testing differences between
peak exercise CI and peak exercise PCWP (𝛽 1 = −1.2, 95% CI
groups were age-adjusted using a linear regression model. Further-
−2.9 to 0.5; R2 = 0.09, P = 0.17). There was no significant corre-
more, the correlations between exercise capacity and invasive haemo-
dynamics were adjusted for beta-blocker use. Pearson’s coefficient of lation between LVSW and LVTMP in CA patients (𝛽 1 = 0.10, 95%
correlation was calculated for normally distributed data, as was Spear- CI −0.04 to 0.25; R2 = 0.11, P = 0.14).
man’s coefficient of correlation for non-normally distributed data. All Findings showed a significantly reduced CO reserve in relation
tests were two-sided. A P-value of <0.05 was considered to indicate to metabolic demands. Hence, the ΔCO/ΔVO2 ratio was lower
a difference of statistical significance. Analyses were performed using in CA subjects than in controls (4.4 ± 1.7 mL blood/mL oxygen vs.
STATA/IC version 13.0 (StataCorp LP, College Station, TX, USA). 6.3 ± 1.6 mL blood/mL oxygen; P = 0.001). Furthermore, the ratio
between ΔSV and ΔOP was 0.9 ± 2.9 mL blood/mL oxygen in CA
patients vs. 3.0 ± 2.2 mL blood/mL oxygen in controls (P < 0.05).
Results At peak exercise, a strong relationship between SV and OP was
Patient characteristics apparent (Figure 3).
In exercise, significantly more pronounced increases in left
Symptomatic heart failure was prevalent in CA patients. Hence,
and right heart filling pressures were found in CA patients
79% of CA patients were of NYHA functional class I or higher
than in controls (Figures 2 and 4). Thus, mean ± SD ΔPCWP
status. Furthermore, CA patients had significantly higher levels of
was 16 ± 6 mmHg vs. 6 ± 5 mmHg, mean ± SD ΔPAP was
N-terminal pro-brain natriuretic peptide (NT-proBNP) and tro-
21 ± 10 mmHg vs. 11 ± 5 mmHg, and mean ± SD ΔRAP was
ponin T than control subjects. CA patients were significantly older
9 ± 5 vs. 0 ± 4 (P-values adjusted for baseline levels, all <0.0001).
than controls (P < 0.0001) (Table 1).
Interestingly, the ratio between SV and PCWP (SV/PCWP) at rest
was significantly lower in CA patients than in controls (P < 0.0001).
Invasive haemodynamic parameters With exercise, the ratio decreased in CA patients (5.3 mL/mmHg
at rest at rest vs. 2.5 mL/mmHg at peak exercise; P < 0.001), whereas no
significant difference was observed in controls (11.2 mL/mmHg vs.
Table 2 displays the results of RHC at rest. Despite higher rest-
10.6 mL/mmHg; P = 0.74) (Figure 2).
ing HR (P < 0.01), CA patients had reduced CI (P < 0.01) as a
result of severely reduced SVI (31 ± 8 mL/m2 vs. 47 ± 11 mL/m2 ; During exercise, SVRI and PVRI were significantly elevated, and
P < 0.0001). CA patients also showed significantly reduced LVSWI PAC severely reduced in CA patients compared with controls.
(Figure 1) and LV-PASW (P < 0.05). CA subjects tended to have After HR adjustment, PAC remained significantly affected in the
impaired VO2 (P = 0.08) and higher mRAP (P = 0.06) than con- CA group (P < 0.0001).
trol subjects. Furthermore, mPAP and mPCWP were significantly At peak exercise, 21 (88%) CA patients had PAP of ≥30 mmHg
higher in the CA group than among controls. Findings showed and 17 (71%) CA patients had PCWP of ≥25 mmHg. Notably, the
mPAP of ≥20 mmHg in 15 (63%) and mPCWP of ≥15 mmHg in increase in right heart pressure was more pronounced than the
12 (50%) CA patients. CA patients receiving loop diuretics had increase in LV filling pressure. Thus, the ΔmRAP/ΔmPCWP ratio
significantly higher mPCWP than CA patients not receiving loop was −0.1 ± 0.7 in the control group, but 0.6 ± 0.4 in CA patients
diuretics [median: 17 mmHg (IQR: 12–22 mmHg) vs. 10 mmHg (P < 0.01). Despite the pronounced increase in right heart filling
(IQR: 6–14 mmHg); P < 0.01]. Both SVRI and PVRI were higher pressures, the RVSWI was comparable between CA subjects and
in CA patients than in controls. Furthermore, CA patients had controls (Figure 1). In the CA-positive group, there was a significant
remarkably lower PAC than controls. inverse relationship between peak exercise RAP and peak exercise
In CA patients, there was no significant correlation between CI (𝛽 1 = −2.29, 95% CI −3.83 to −0.75; R2 = 0.31, P < 0.001).
LVTMP and LVSWI [𝛽 1 = −0.12, 95% confidence interval (CI) In CA patients, there was an inverse relationship between
−0.31 to 0.07; R2 = 0.07, P = 0.21]. troponin T and CI at peak exercise (𝛽 1 = −16, 95% CI −23 to −9;
R2 = 0.56, P < 0.0001). Furthermore, troponin T correlated weakly
with peak exercise RAP (𝛽1 = 2.3, 95% CI 0.2–4.5; R2 = 0.22,
Invasive haemodynamic parameters P < 0.05), but not with peak exercise PCWP (𝛽 1 = 1.00, 95% CI
at peak exercise −1.5 to 3.5; R2 = 0.04, P = 0.41). Results showed a weak correlation
Patients with CA had lower VO2 than controls (15 ± 6 mL/min/kg between NT-proBNP and peak exercise CI (𝛽 1 = −1114, 95% CI
bwt vs. 33 ± 7 mL/min/kg bwt; P < 0.0001). −1962 to −267; R2 = 0.27, P < 0.05). No significant correlations

© 2017 The Authors

European Journal of Heart Failure © 2017 European Society of Cardiology
4 T.S. Clemmensen et al.

Table 1 Patient characteristics and cardiac amyloid burden in subjects with (Group A) and without (Group B) cardiac
amyloidosis

Group A Group B P-value

(n = 24) (n = 20)
...........................................................................................................................................
Male 83% 75% 0.50
Weight, kg, mean ± SD 77 ± 13 76 ± 16 0.74
BSA, m2 , mean ± SD 1.9 ± 0.2 1.9 ± 0.2 0.93
Age, years, mean ± SD 67 ± 15 47 ± 12 <0.0001
NYHA functional class > 79% 0% <0.0001
Diabetes 0% – –
Hypertension 22% – –
Medication
Statins 17% – –
ACEi/ARB 17% – –
Beta-blockers 29% – –
Calcium channel blocker 0% – –
Furosemide or bumetanide 67% – –
Thiazide 8% – –
Spironolactone 17% – –
Biochemistry
Lactate at peak exercise, mmol/L, median (IQR) 4.9 (3.0–6.2) 10 (8.5–12.2) <0.0001a
Creatinine, μmol/L, median (IQR) 101 (83–132) 75 (69–86) <0.001b
Haemoglobin, mmol/L, mean ± SD 8.4 ± 1.0 9.0 ± 0.9 0.08
Troponin T, ng/L, median (IQR) 65 (25–90) 6 (6–6) <0.0001a
NT-proBNP, ng/L, median (IQR) 2762 (734–4359) 46 (34–78) <0.0001b

ACEi, angiotensin-converting enzyme inhibitor; ARB, angiotensin receptor blocker; BSA, body surface area; IQR, interquartile range; NT-proBNP, N-terminal pro-brain
natriuretic peptide; NYHA, New York Heart Association; SD, standard deviation.
a P < 0.01 after age adjustment.
b P < 0.05 after age adjustment.

between NT-proBNP and peak exercise RAP (P = 0.06) or peak patients. By contrast, no correlation was seen between 3-D LV
...........................................................................

exercise PCWP (P = 0.10) were observed. mass and right or left heart filling pressures.
Patients in the CA group had lower 2-D LV ejection fraction
(LVEF) (P < 0.05) and lower 2-D LV EDV (LVEDV) than the controls
Relationship between exercise capacity (P < 0.05). Six patients with CA and none of the controls had
and haemodynamic parameters 2-D LVEF of <45%. With exercise, 2-D LVEF increased less in
Table 3 shows correlation coefficients between peak exercise VO2 CA patients than in controls (mean ± SD ΔLVEF: 2.1 ± 8.4% vs.
and haemodynamic parameters in CA patients (Group A). There 10.0 ± 6.3%; P < 0.01). At peak exercise, 2-D LV end systolic
was a very strong relationship between peak exercise VO2 and volume was significantly higher in CA patients than in controls
peak exercise CI (Figure 3). Furthermore, significant relationships [median: 36 mL (IQR: 24–55 mL) vs. 27 mL (IQR: 24–30 mL);
between peak exercise VO2 and PAC, LVSWI and RV filling pressure P < 0.05], whereas no significant difference in LVEDV was observed
were noted. These correlations remained significant after adjust- [87 mL (IQR: 70–98 mL) vs. 96 mL (IQR: 85–109 mL); P = 0.21].
ments for age and for use of beta-blockers. As expected, CA patients had severely lower long-axis function
by tissue tracking than controls. Tissue tracking values were
strongly correlated with both resting and peak exercise CI (rest:
Relationship between non-invasive 𝛽 1 = 3.7, 95% CI 1.9–5.5; R2 = 0.46, P < 0.0001; peak: 𝛽 1 = 1.3, 95%
and invasive myocardial performance CI 1.0–1.7; R2 = 0.72, P < 0.0001) and PAC (rest: 𝛽 1 = 1.0, 95%
Table 4 shows diastolic and systolic myocardial performance by CI 0.5–1.5; R2 = 0.45, P < 0.0001; peak: 𝛽 1 = 1.5, 95% CI 0.8–2.1;
echocardiography. As expected, LV mass was significantly higher in R2 = 0.55, P < 0.0001). By contrast, no correlation between tissue
CA patients than in the control group as assessed by 3-D echocar- tracking and right or left heart filling pressures was seen.
diography [median: 125 g/m2 (IQR: 111–155 g/m2 ) vs. 80 g/m2 Diastolic performance clearly indicated restrictive physiology in
(IQR: 71–84 g/m2 ); P < 0.0001]. Results showed 3-D LV mass to the majority of CA patients. Hence, 19 of 24 (79%) CA patients had
be significantly correlated with peak exercise CI (𝛽 1 = −9.6, 95% diastolic dysfunction. The E/E’ ratio was the only echocardiographic
CI −15.1 to −4.0; R2 = 0.39, P < 0.01) and peak exercise PAC parameter to be correlated with peak exercise PCWP in CA
(𝛽 1 = −12.2, 95% CI −18.6 to −5.8; R2 = 0.44, P < 0.01) in CA patients (𝛽 1 = 0.9, 95% CI 0.3–1.5; R2 = 0.36, P < 0.01).

© 2017 The Authors

European Journal of Heart Failure © 2017 European Society of Cardiology
Exercise haemodynamics in cardiac amyloidosis 5

Table 2 Haemodynamic parameters at rest and at peak exercise in subjects with (Group A) and without (Group B)
cardiac amyloidosis

At rest At peak exercise

.................................................. ...................................................
Group A Group B P-value Group A Group B P-value
(n = 24) (n = 20) (n = 24) (n = 20)
...........................................................................................................................................
Watt, median (IQR) 75 (30–100) 175 (150–225) <0.0001a
Haemodynamics
MAP, mmHg, mean ± SD 91 ± 10 94 ± 11 0.35 104 ± 28 128 ± 23 <0.01c
HR, b.p.m., median (IQR) 76 (64–84) 64 (56–69) <0.01b 125 (105–141) 163 (137–173) <0.001
AV-diff, %, median (IQR) 36 (26–40) 23 (21–26) <0.0001b 69 (60–82) 66 (62–71) 0.34
SVRI, dynes*s*cm−5 *m2 , median (IQR) 3084 (2440–3727) 2411 (1958–3079) <0.05 1662 (1234–2201) 1082 (1014–1127) <0.001c
CI, L/min/m2 , mean ± SD 2.3 ± 0.6 2.9 ± 0.7 <0.01 4.3 ± 2.0 9.8 ± 1.8 <0.0001a
VO2 , mL/min/kg, mean ± SD 3.9 ± 0.8 4.4 ± 0.9 0.08 15.0 ± 6.0 33.0 ± 7.0 <0.0001a
mRAP, mmHg, median (IQR) 6 (5–8) 5 (4–6) 0.06 14 (10–19) 3 (2–6) <0.0001b
mPAP, mmHg, median (IQR) 25 (18–29) 15 (14–17) <0.001b 45 (40–52) 26 (22–30) <0.0001a
mPCWP, mmHg, median (IQR) 15 (11–20) 8 (7–10) <0.001b 32 (25–35) 14 (11–18) <0.0001a
mRAP/mPCWP, median (IQR) 0.5 (0.3–0.6) 0.6 (0.5–0.7) 0.11 0.5 (0.3–0.6) 0.2 (0.1–0.5) <0.05c
LVTMP, mmHg, median (IQR) 8 (4–12) 4 (3–4) <0.01b 17 (10–20) 12 (7–13) <0.05c
PVRI, dynes*s*cm−5 *m2 , median (IQR) 288 (194–528) 173 (146–215) <0.01 243 (194–398) 101 (79–119) <0.0001c
PAC, mL/mmHg, median (IQR) 2.9 (2.1–4.5) 7.5 (5.7–10.4) <0.0001b 2.2 (1.6–2.8) 7.0 (6.4–8.6) <0.0001a
LV-PASW, median (IQR) 56 (43–76) 74 (65–82) <0.05 70 (55–112) 157 (135–172) <0.0001a

AV-diff, arterial–venous saturation difference; CI, cardiac index; HR, heart rate; IQR, interquartile range; LV-PASW, left ventricular preload adjusted stroke work; LVTMP, left
ventricular transmural pressure; MAP, mean arterial pressure; mPAP, mean pulmonary arterial pressure; mPCWP, mean pulmonary capillary wedge pressure; mRAP, mean right
atrial pressure; PAC, pulmonary arterial compliance; PVRI, pulmonary vascular resistance index; SD, standard deviation SVRI, systemic vascular resistance index; VO2 , oxygen
consumption.
a P < 0.0001 after age adjustment.
b P < 0.01 after age adjustment.
c P < 0.05 after age adjustment.

(A) (B)
150 Rest Peak exercise 40 Rest Peak exercise
P < 0.0001
Right ventricular stroke work index, g*m/m2/beat
Left ventricular stroke work index, g*m/m2/beat

30
P = 0.13
100

P < 0.0001

20

P = 0.90
50

10

0 0
Controls CA patients Controls CA patients Controls CA patients Controls CA patients

Figure 1 (A) Left ventricular and (B) right ventricular stroke work indices in patients with cardiac amyloidosis (CA) and control subjects at
rest and at peak exercise. P < 0.0001 after age adjustment.

© 2017 The Authors

European Journal of Heart Failure © 2017 European Society of Cardiology
6 T.S. Clemmensen et al.

(A) (B) (C)

140 40 20

Stroke volume/PCWP, mL/mmHg

120 30 15
Stroke volume, mL

PCWP, mmHg
100
20 10

80
10 5

60
0 0
Rest 0 Watt 20/50 Watt Peak Recovery Rest 0 Watt 20/50 Watt Peak Recovery Rest 0 Watt 20/50 Watt Peak Recovery

Group B: Controls, n = 20 Group B: Controls, n = 20 Group B: Controls, n = 20

Group A: CA positive, n = 24 Group A: CA positive, n = 24 Group A: CA positive, n = 24

Figure 2 (A) Stroke volume, (B) pulmonary capillary wedge pressure (PCWP), and (C) stroke volume/PCWP during exercise in subjects with
(Group A) and without (Group B) cardiac amyloidosis (CA). Bars indicate 95% confidence intervals.

(A) (B)
200 All individuals: R2 = 0.77, P < 0.0001 15 All individuals: R2 = 0.84, P < 0.0001

Peak exercise cardiac index, L/min/m2

Peak exercise stroke volume, mL/beat

CA patients: R2 = 0.67, P < 0.0001 CA patients: R2 = 0.76, P < 0.0001

150
10

100

5
50

0 –0
5 10 15 20 10 20 30 40 50
Peak exercise oxygen pulse (mL/beat) Peak exercise oxygen consumption (VO2, L/min/kg)

Group A: CA patients Group A: CA patients

Group B: Control population Group B: Control population

Figure 3 (A) Peak exercise stroke volume and peak exercise oxygen pulse, and (B) peak exercise cardiac index and peak exercise oxygen
consumption (VO2 ) in subjects with (Group A) and without (Group B) cardiac amyloidosis (CA).

Discussion afterload or decrease in contractility. Thus, the observed lower

.......................................................

SV in CA patients may result from lower EDV, increased SVRI

This is the first study to evaluate the haemodynamic mechanisms
and/or reduced contractile reserve. LV stroke work, which is
underlying impaired exercise capacity in CA patients. Findings
independent of afterload, increased much less in CA, which indi-
showed that CA patients had profound myocardial reserve defi-
cates that the increased afterload is not the main contributor to
ciency, which was significantly related to their poorer exercise
reduced SV in CA. At peak exercise, LV-PASW was significantly
capacity. Furthermore, they had reduced PAC and severely
lower in CA patients than in controls. Furthermore, there was no
increased left and especially right heart filling pressures during
exercise. correlation between LVTMP and LVSWI at rest or during exercise,
Traditionally, CA is considered a predominantly diastolic heart which suggests that preload reduction is not responsible for the
failure disease caused by increased myocardial stiffness and restric- reduced SV in CA. Therefore, impaired contractility is most likely
tive physiology. Hence, several studies found that ATTRwt is a to be responsible for the observed SV reduction. The inotropic
relatively common underlying mechanism in heart failure with reserve deficiency in CA patients was accompanied by forward
preserved ejection fraction (HFpEF).19,20 Like HFpEF patients in heart failure. Hence, CA patients had reduced ΔCO/ΔVO2 . The
general, CA patients in the present study had preserved LVEF. combined systolic and diastolic impairment and low myocardial
However, the current results clearly demonstrate a profound reserve demonstrated in this study are remarkably similar to those
lack of myocardial contractile reserve in CA patients, which was described in HFpEF.14,21,22 Furthermore, exercise VO2 in HFpEF
significantly associated with lack of physical capacity and peak shows a relationship to invasive haemodynamics similar to that
exercise VO2 . SV drops with a decrease in preload, increase in demonstrated here in CA patients.22

© 2017 The Authors

European Journal of Heart Failure © 2017 European Society of Cardiology
Exercise haemodynamics in cardiac amyloidosis 7

(A) (B)
20 50

Pulmonary arterial pressure, mmHg

Right atrial pressure, mmHg
15 40

10 30

5 20

0 10

Rest 0 Watt 20/50 Watt Peak Recovery Rest 0 Watt 20/50 Watt Peak Recovery

Group B, n = 20 Group B, n = 20
Group A, n = 24 Group A, n = 24

Figure 4 (A) Mean right atrial pressure and (B) mean pulmonary arterial pressure in subjects with (Group A) and without (Group B) cardiac
amyloidosis. Bars indicate 95% confidence intervals.

The present study shows that exercise testing unmasked 66%

...........................................................................................................
Table 3 Univariate and age-adjusted analysis of the
(six of nine) patients with normal resting filling pressure. This
relationship between maximal oxygen consumption
emphasizes that assessment of haemodynamics during resting con-
(VO2 ) and haemodynamics in cardiac amyloidosis
ditions falls short in evaluating myocardial systolic and diastolic per-
patients
formance. Exercise testing should therefore be considered during
haemodynamic assessment in patients with CA, as recommended
VO2
in patients with HFpEF.23,24 Univariate P-value
Among CA patients, SVRI during exercise was elevated com- ...................
pared with controls. This finding probably demonstrates a 𝜷1 R2
................................................................
mechanism that compensates for the observed decreases in LV Peak watt level 0.12 0.84 <0.0001a,b
SV and stroke work performed by the left ventricle to maintain a HR, peak exercise 0.13 0.44 <0.001c,d
partly adequate MAP, which is actually decreased in CA patients. SVRI, peak exercise −0.01 0.26 <0.05
These findings are of great importance as these haemodynamic CI, peak exercise 2.4 0.76 <0.0001a,b
alterations may contribute to symptoms commonly encountered SV, peak exercise 0.13 0.38 <0.01c,d
in CA patients, such as dizziness, syncope and fatigue. LVSWI, peak exercise 0.03 0.45 <0.01c,d
The severely reduced PAC and elevated right-sided pressures mRAP, peak exercise −0.36 0.30 <0.01c,d
are other key features of CA patients’ exercise haemodynamics. mPAP, peak exercise −0.06 0.01 0.61
PAC is believed to be a measure of pulmonary arterial distensibility mPCWP, peak exercise −0.20 0.08 0.19
PAC, peak exercise 2.3 0.45 <0.001c,d
expressing the absolute change in volume for a given change in
pressure. PAC has been shown to possess important prognostic CI, cardiac index; HR, heart rate; LVSWI, left ventricular stroke work index;
value in both primary pulmonary hypertension and left-sided mPAP, mean pulmonary arterial pressure; mPCWP, mean pulmonary capillary
heart failure.25 – 28 The demonstrated impaired pulmonary com- wedge pressure; mRAP, mean right atrial pressure; PAC, pulmonary arterial
compliance; SV, stroke volume; SVRI, systemic vascular resistance index.
pliance may be attributable to backward failure, which may be a P < 0.0001 after age adjustment.
b P < 0.0001 after beta-blocker use adjustment.
deteriorated by pericardial restraint as a result of elevated total
c P < 0.05 after age adjustment.
heart volume in CA patients. However, impaired PAC may also d P < 0.05 after beta-blocker use adjustment.
indicate the presence of amyloid deposits in the pulmonary arterial
vessels. The latter explanation was supported by the relatively
higher exercise-induced increase in RAP than PCWP, which is in alteration, although deterioration of RV myocardial performance
accordance with the elevated pulmonary pressure in CA patients. is common in CA patients. Elevated RAP4 and RV dysfunction30,31
However, no clinically relevant increase in PVRI was demonstrated have both shown important independent prognostic value in CA
at either rest or peak exercise. The relatively high impedance of the patients, whereas the prognostic value of PAC must be clarified in
pulmonary vascular bed makes the right ventricle very dependent future studies.
on its loading conditions.29 Therefore, reduced PAC has graver In the present study, 3-D echocardiography was used to assess
consequences for RV afterload and entails an increased risk for LV mass as a marker of interstitial amyloid burden. Traditionally,
later RV failure. The poor RV output demonstrated in the present 2-D M-mode LV mass assessment has been used to determine
study may therefore be a consequence of pulmonary vascular the diagnosis and severity of CA. However, the ability of 2-D

© 2017 The Authors

European Journal of Heart Failure © 2017 European Society of Cardiology
8 T.S. Clemmensen et al.

Table 4 Echocardiographic parameters at rest in subjects with (Group A) and without (Group B) cardiac amyloidosis

Group A Group B P-value

(n = 24) (n = 20)
...........................................................................................................................................
3-D LV mass, g/m2 , median (IQR) 125 (111–155) 80 (71–84) <0.0001a
Systolic function, median (IQR)
2-D EF, % 59 (45–63) 63 (60–67) <0.05
2-D EDV, mL 93 (73–108) 116 (100–1131) <0.05
2-D ESV, mL 39 (31–60) 43 (33–47) 0.63
2-D Tissue tracking, mm 5 (4–8) 14 (13–16) <0.0001
Diastolic function, median (IQR)
LA volume, mL/m2 40 (31–54) 23 (19–33) <0.001c
E, cm/s 89 (79–15) 70 (62–78) <0.001b
E/A ratio 1.8 (1.3–3.7) 1.6 (1.3–2.1) 0.17
E’ lateral, cm/s 4.4 (3.5–6.5) 9.8 (9.1–12.4) <0.0001a
E/E’ ratio 20.8 (15.7–29.0) 7.3 (5.6–8.2) <0.0001a
E-deceleration time, ms 161 (137–190) 205 (166–244) <0.05b
TR gradient, mmHg 29 (20–37) 23 (19–25) <0.05
ASE diastolic classification, n (%)
Normal 2 (8) 14 (56) <0.0001a
Intermediate 3 (13) 5 (25) 0.29
Diastolic dysfunction 19 (79) 1 (5) <0.0001a

3-D, three-dimensional; ASE, American Society of Echocardiography; EDV, end-diastolic volume; EF, ejection fraction; ESV, end-systolic volume; IQR, interquartile range; LA,
left atrium; TR, tricuspid regurgitation.
a P < 0.0001 after age adjustment.
b P < 0.01 after age adjustment.
c P < 0.05 after age adjustment.

assessment to detect asymmetric LV hypertrophy is limited. In Limitations

...................................................................................

contrast, magnetic resonance imaging (MRI) to assess the LV mass

There are a number of limitations to this study. It reflects the
index enjoys high diagnostic accuracy and outperforms tradi-
experience of a single centre with a relatively small cohort of
tional diagnostic criteria.32 However, the use of MRI is costly and
patients. The age difference between CA patients and controls
time-consuming and is often associated with patient discomfort
is a major confounder. Therefore, all P-values were adjusted
as a result of claustrophobia and anxiety. In contrast, LV mass
for age. The difference in beta-blocker use may be an impor-
assessment by 3-D echocardiography is simple and easily available tant confounder. However, because the number of patients
and can be performed at the bedside. Moreover, it compares studied was limited, the present group was unable to adjust
favourably with MRI.33,34 In the current study, LV mass index was for all potential confounders. Thus, relationships between
found to be a good indirect marker of the amyloid burden and a exercise capacity and invasive haemodynamics were adjusted
significant relationship between invasive systolic haemodynamic only for beta-blocker use. Haemodynamic assessment can
parameters and PAC was observed. Likewise, long-axis displace- be challenging during exercise as a result of respiratory and
ment by tissue tracking was strongly associated with invasive cardiac motion.
systolic performance. However, neither 3-D LV mass nor tissue
tracking correlated with invasive filling pressures. Only E/E’ weakly
correlated with peak exercise PCWP. Therefore, the clinical utility Conclusions
of resting echocardiography for filling pressure prediction during
Exercise haemodynamics in CA are characterized by impairment of
exercise is poor. A recent study demonstrated great clinical utility inotropic myocardial reserve and PAC, along with increased right-
by exercise tissue Doppler echocardiography for filling pressure and left-sided pressures. The lack of PAC and inotropic myocardial
prediction in HFpEF patients.35 As the haemodynamics response to reserve strongly correlated with the lack of exercise capacity and
exercise in CA and HFpEF shows remarkable similarities, exercise VO2 .
tissue Doppler echocardiography may also prove beneficial in
diastolic performance evaluation in CA patients. A well known
echocardiographic feature of CA is the lack of longitudinal myocar-
Acknowledgements
dial deformation.36 Therefore, future studies should evaluate the Nurses and physicians at the outpatient clinic and cardiac labora-
relationship between myocardial deformation imaging and invasive tory are thanked for their invaluable assistance with right heart
haemodynamics in CA. catheterization and stress testing.

© 2017 The Authors

European Journal of Heart Failure © 2017 European Society of Cardiology
Exercise haemodynamics in cardiac amyloidosis 9

Funding amyloid deposition in patients with heart failure and preserved ejection fraction.

.....................................................................................................................................................
JACC Heart Fail 2014;2:113–122.
Health Research Fund of Central Denmark Region and the Danish 20. Gonzalez-Lopez E, Gallego-Delgado M, Guzzo-Merello G, de Haro-Del Moral FJ,
Cobo-Marcos M, Robles C, Bornstein B, Salas C, Lara-Pezzi E, Alonso-Pulpon
Heart Association. L, Garcia-Pavia P. Wild-type transthyretin amyloidosis as a cause of heart failure
Conflict of interest: none declared. with preserved ejection fraction. Eur Heart J 2015;36:2585–2594.
21. Borlaug BA, Olson TP, Lam CS, Flood KS, Lerman A, Johnson BD, Redfield MM.
Global cardiovascular reserve dysfunction in heart failure with preserved ejection
References fraction. J Am Coll Cardiol 2010;56:845–854.
1. Quarta CC, Kruger JL, Falk RH. Cardiac amyloidosis. Circulation 22. Borlaug BA, Kane GC, Melenovsky V, Olson TP. Abnormal right
2012;126:e178–182. ventricular–pulmonary artery coupling with exercise in heart failure with
2. Falk RH. Cardiac amyloidosis: a treatable disease, often overlooked. Circulation preserved ejection fraction. Eur Heart J 2016;37:3293–3302.
2011;124:1079–1085. 23. Borlaug BA, Nishimura RA, Sorajja P, Lam CS, Redfield MM. Exercise hemody-
3. Falk RH, Comenzo RL, Skinner M. The systemic amyloidoses. N Engl J Med namics enhance diagnosis of early heart failure with preserved ejection fraction.
1997;337:898–909. Circ Heart Fail 2010;3:588–595.
4. Russo C, Green P, Maurer M. The prognostic significance of central hemodynam- 24. Dorfs S, Zeh W, Hochholzer W, Jander N, Kienzle RP, Pieske B, Neumann FJ.
ics in patients with cardiac amyloidosis. Amyloid 2013;20:199–203. Pulmonary capillary wedge pressure during exercise and long-term mortality in
5. Granstam SO, Rosengren S, Vedin O, Kero T, Sorensen J, Carlson K, Flachskampf patients with suspected heart failure with preserved ejection fraction. Eur Heart
FA, Wikström G. Evaluation of patients with cardiac amyloidosis using echocar- J 2014;35:3103–3112.
diography, ECG and right heart catheterization. Amyloid 2013;20:27–33. 25. Miller WL, Grill DE, Borlaug BA. Clinical features, hemodynamics, and out-
6. Rapezzi C, Merlini G, Quarta CC, Riva L, Longhi S, Leone O, Salvi F, Ciliberti comes of pulmonary hypertension due to chronic heart failure with reduced
P, Pastorelli F, Biagini E, Coccolo F, Cooke RM, Bacchi-Reggiani L, Sangiorgi D, ejection fraction: pulmonary hypertension and heart failure. JACC Heart Fail
Ferlini A, Cavo M, Zamagni E, Fonte ML, Palladini G, Salinaro F, Musca F, Obici 2013;1:290–299.
L, Branzi A, Perlini S. Systemic cardiac amyloidoses: disease profiles and clinical 26. Pellegrini P, Rossi A, Pasotti M, Raineri C, Cicoira M, Bonapace S, Dini FL,
courses of the 3 main types. Circulation 2009;120:1203–1212. Temporelli PL, Vassanelli C, Vanderpool R, Naeije R, Ghio S. Prognostic relevance
7. Trikas A, Rallidis L, Hawkins P, Oakley CM, Nihoyannopoulos P. Comparison of pulmonary arterial compliance in patients with chronic heart failure. Chest
of usefulness between exercise capacity and echocardiographic indexes of left 2014;145:1064–1070.
ventricular function in cardiac amyloidosis. Am J Cardiol 1999;84:1049–1054. 27. Scarsini R, Prioli MA, Milano EG, Castellani C, Pesarini G, Assael BM, Vassanelli
8. Antoni G, Lubberink M, Estrada S, Axelsson J, Carlson K, Lindsjo L, Kero T, C, Ribichini SL. Hemodynamic predictors of long term survival in end stage cystic
Långström B, Granstam SO, Rosengren S, Vedin O, Wassberg C, Wikström G, fibrosis. Int J Cardiol 2016;202:221–225.
Westermark P, Sörensen J. In vivo visualization of amyloid deposits in the heart 28. Thenappan T, Prins KW, Pritzker MR, Scandurra J, Volmers K, Weir EK. The
with 11 C-PIB and PET. J Nucl Med 2013;54:213–220. critical role of pulmonary arterial compliance in pulmonary hypertension. Ann
9. Lee SP, Lee ES, Choi H, Im HJ, Koh Y, Lee MH, Kwon JH, Paeng JC, Kim HK, Am Thorac Soc 2016;13:276–284.
Cheon GJ, Kim YJ, Kim I, Yoon SS, Seo JW, Sohn DW. 11 C-Pittsburgh B PET 29. Kjaergaard J, Petersen CL, Kjaer A, Schaadt BK, Oh JK, Hassager C. Evaluation of
imaging in cardiac amyloidosis. JACC Cardiovasc Imaging 2015;8:50–59. right ventricular volume and function by 2D and 3D echocardiography compared
10. Pilebro B, Arvidsson S, Lindqvist P, Sundstrom T, Westermark P, Antoni G, Suhr to MRI. Eur J Echocardiogr 2006;7:430–438.
O, Sörensen J. Positron emission tomography (PET) utilizing Pittsburgh com- 30. Bellavia D, Pellikka PA, Dispenzieri A, Scott CG, Al-Zahrani GB, Grogan M,
pound B (PIB) for detection of amyloid heart deposits in hereditary transthyretin Pitrolo F, Oh JK, Miller FA Jr. Comparison of right ventricular longitudinal
amyloidosis (ATTR). J Nucl Cardiol 2016;doi: 10.1007/s12350-016-0638-5. strain imaging, tricuspid annular plane systolic excursion, and cardiac biomarkers
11. Andersen MJ, Borlaug BA. Invasive hemodynamic characterization of heart failure for early diagnosis of cardiac involvement and risk stratification in primary
with preserved ejection fraction. Heart Fail Clin 2014;10:435–444. systematic (AL) amyloidosis: a 5-year cohort study. Eur Heart J Cardiovasc Imaging
12. van Empel VP, Kaye DM, Borlaug BA. Effects of healthy aging on the cardiopul- 2012;13:680–689.
monary hemodynamic response to exercise. Am J Cardiol 2014;114:131–135. 31. Cappelli F, Porciani MC, Bergesio F, Perlini S, Attana P, Moggi Pignone A,
13. Kovacs G, Berghold A, Scheidl S, Olschewski H. Pulmonary arterial pressure Salinaro F, Musca F, Padeletti L, Perfetto F. Right ventricular function in AL
during rest and exercise in healthy subjects: a systematic review. Eur Respir J amyloidosis: characteristics and prognostic implication. Eur Heart J Cardiovasc
2009;34:888–894. Imaging 2012;13:416–422.
14. Abudiab MM, Redfield MM, Melenovsky V, Olson TP, Kass DA, Johnson BD, 32. Pozo E, Kanwar A, Deochand R, Castellano JM, Naib T, Pazos-Lopez P, Osman
Borlaug BA. Cardiac output response to exercise in relation to metabolic demand K, Cham M, Narula J, Fuster V, Sanz J. Cardiac magnetic resonance evalua-
in heart failure with preserved ejection fraction. Eur J Heart Fail 2013;15:776–785. tion of left ventricular remodelling distribution in cardiac amyloidosis. Heart
15. Nagueh SF, Smiseth OA, Appleton CP, Byrd BF III, Dokainish H, Edvardsen T, 2014;100:1688–1695.
Flachskampf FA, Gillebert TC, Klein AL, Lancellotti P, Marino P, Oh JK, Popescu 33. Jenkins C, Bricknell K, Hanekom L, Marwick TH. Reproducibility and accu-
BA, Waggoner AD. Recommendations for the evaluation of left ventricular racy of echocardiographic measurements of left ventricular parameters using
diastolic function by echocardiography: an update from the American Society of real-time three-dimensional echocardiography. J Am Coll Cardiol 2004;44:
Echocardiography and the European Association of Cardiovascular Imaging. Eur 878–886.
Heart J Cardiovasc Imaging 2016;29:277–314. 34. Mor-Avi V, Sugeng L, Weinert L, MacEneaney P, Caiani EG, Koch R, Salgo IS, Lang
16. Pan C, Hoffmann R, Kuhl H, Severin E, Franke A, Hanrath P. Tissue tracking allows RM. Fast measurement of left ventricular mass with real-time three-dimensional
rapid and accurate visual evaluation of left ventricular function. Eur J Echocardiogr echocardiography: comparison with magnetic resonance imaging. Circulation
2001;2:197–202. 2004;110:1814–1818.
17. Andersen NH, Poulsen SH. Evaluation of the longitudinal contraction of the left 35. Obokata M, Kane GC, Reddy YN, Olson TP, Melenovsky V, Borlaug BA. The
ventricle in normal subjects by Doppler tissue tracking and strain rate. J Am Soc role of diastolic stress testing in the evaluation for HFpEF: a simultaneous
Echocardiogr 2003;16:716–723. invasive-echocardiographic study. Circulation 2017;135:825–838.
18. Wolsk E, Bakkestrøm R, Thomsen JH, Balling L, Andersen MJ, Dahl JS, Hassager C, 36. Ternacle J, Bodez D, Guellich A, Audureau E, Rappeneau S, Lim P, Radu C, Guen-
Møller JE, Gustafsson F. The influence of age on hemodynamic parameters during douz S, Couetil JP, Benhaiem N, Hittinger L, Dubois-Randé JL, Plante-Bordeneuve
rest and exercise in healthy individuals. JACC Heart Fail 2017;5(5):337–346. V, Mohty D, Deux JF, Damy T. Causes and consequences of longitudinal LV dys-
19. Mohammed SF, Mirzoyev SA, Edwards WD, Dogan A, Grogan DR, Dunlay SM, function assessed by 2D strain echocardiography in cardiac amyloidosis. JACC
Roger VL, Gertz MA, Dispenzieri A, Zeldenrust SR, Redfield MM. Left ventricular Cardiovasc Imaging 2016;9:126–138.

© 2017 The Authors

European Journal of Heart Failure © 2017 European Society of Cardiology