Modelling the lifetime cost-effectiveness of catheter ablation for atrial fibrillation with heart failure

Abstract

Objectives
Assessing the cost-effectiveness credentials of this intervention in patients with concomitant atrial fibrillation (AF) and heart failure (HF) compared with usual medical therapy.

Design
A Markov model comprising two health states (ie, alive or dead) was constructed. The transition probabilities were directly derived from published Kaplan-Meier curves of the pivotal randomised controlled trial and extrapolated over the cohort’s lifetime using recommended methods. Costs of catheter ablation, outpatient consultations, hospitalisation, medications and examinations were included. Resource use and unit costs were sourced from government websites or published literature. A lifetime horizon and a healthcare system perspective were taken. All costs and benefits were discounted at 3% annually. Deterministic (DSA) and probabilistic sensitivity analyses (PSA) were run around the key model parameters to test the robustness of the base case results.

Participants
A hypothetical Australian cohort of patients with concomitant AF and HF who are resistant to antiarrhythmic treatment.

Interventions
Catheter ablation versus medical therapy.

Results
The catheter ablation was associated with a cost of $A44 377 per person, in comparison to $A28 506 for the medical therapy alone over a lifetime. Catheter ablation contributed to 4.58 quality-adjusted life years (QALYs) and 6.99 LY gains compared with 4.30 QALYs and 6.53 LY gains, respectively, in the medical therapy arm. The incremental cost-effectiveness ratio was $A55 942/QALY or $A35 020/LY. The DSA showed that results were highly sensitive to costs of ablation and time horizon. The PSA yielded very consistent results with the base case.

Conclusions
Offering catheter ablation procedure to patients with systematic paroxysmal or persistent AF who failed to respond to antiarrhythmic drugs was associated with higher costs, greater benefits. When compared with medical therapy alone, this intervention is not cost-effective from an Australia healthcare system perspective.

Full text

1
GaoL, MoodieM. BMJ Open 2019;9:e031033. doi:10.1136/bmjopen-2019-031033
Open access
Modelling the lifetime cost-
effectiveness of catheter ablation for
atrial brillation with heart failure
Lan Gao,
Marj Moodie
To cite: GaoL, MoodieM.
Modelling the lifetime cost-
effectiveness of catheter
ablation for atrial brillation
with heart failure. BMJ Open
2019;9:e031033. doi:10.1136/
bmjopen-2019-031033
►Prepublication history and
additional material for this
paper are available online. To
view these les, please visit
the journal online (http:// dx. doi.
org/ 10. 1136/ bmjopen- 2019-
031033).
Received 12 April 2019
Revised 01 July 2019
Accepted 12 July 2019
Deakin University, Faculty of
Health, Institute for Health
Transformation, Deakin Health
Economics, Geelong, Victoria,
Australia
Correspondence to
Dr Lan Gao;
lan. gao@ deakin. edu. au
Original research
© Author(s) (or their
employer(s)) 2019. Re-use
permitted under CC BY-NC. No
commercial re-use. See rights
and permissions. Published by
BMJ.
Strengths and limitations of this study
►This is the rst study that used the evidence from re-
cently published randomised controlled trial (RCT) to
assess the cost-effectiveness of catheter ablation in
treating patients concomitant with atrial brillation
and heart failure.
►The reconstructed individual patient data (IPD) was
derived from published Kaplan-Meier curve from the
RCT, which were used to derive the transition proba-
bilities in the Markov model.
►Extensive sensitivity analyses were undertaken to
test the robustness of the results, including consid-
eration of different parametric survival models to
extrapolate the survival observed over the trial.
►The reconstruction of the IPD is only a maximum ap-
proximation of the real data, but the algorithm used
to derive the IPD is considered reliable.
ABSTRACT
Objectives Assessing the cost-effectiveness credentials
of this intervention in patients with concomitant atrial
brillation (AF) and heart failure (HF) compared with usual
medical therapy.
Design A Markov model comprising two health states (ie,
alive or dead) was constructed. The transition probabilities
were directly derived from published Kaplan-Meier curves
of the pivotal randomised controlled trial and extrapolated
over the cohort’s lifetime using recommended methods.
Costs of catheter ablation, outpatient consultations,
hospitalisation, medications and examinations were
included. Resource use and unit costs were sourced from
government websites or published literature. A lifetime
horizon and a healthcare system perspective were taken.
All costs and benets were discounted at 3% annually.
Deterministic (DSA) and probabilistic sensitivity analyses
(PSA) were run around the key model parameters to test
the robustness of the base case results.
Participants A hypothetical Australian cohort of patients
with concomitant AF and HF who are resistant to
antiarrhythmic treatment.
Interventions Catheter ablation versus medical therapy.
Results The catheter ablation was associated with a cost
of $A44 377 per person, in comparison to $A28 506 for
the medical therapy alone over a lifetime. Catheter ablation
contributed to 4.58 quality-adjusted life years (QALYs)
and 6.99 LY gains compared with 4.30 QALYs and 6.53
LY gains, respectively, in the medical therapy arm. The
incremental cost-effectiveness ratio was $A55 942/QALY
or $A35 020/LY. The DSA showed that results were highly
sensitive to costs of ablation and time horizon. The PSA
yielded very consistent results with the base case.
Conclusions Offering catheter ablation procedure to
patients with systematic paroxysmal or persistent AF who
failed to respond to antiarrhythmic drugs was associated
with higher costs, greater benets. When compared
with medical therapy alone, this intervention is not cost-
effective from an Australia healthcare system perspective.
INTRODUCTION
Chronic heart failure (HF) and atrial fibril-
lation (AF) are common conditions that
contribute significantly to the risk of death
worldwide. Both conditions are becoming
increasingly prevalent and resulting in spiral-
ling costs to healthcare systems internation-
ally.1–3 The incidence of AF is predicted to
double over the next 20 years.4 5 HF is the
leading cause of hospitalisation among adults
aged over 65 years of age with more than 41
000 people hospitalised annually in Australia.6
Despite dramatic improvement in outcomes
in patients treated with medical therapy,
more than 50% of patients with HF are rehos-
pitalised within 6 months of discharge,7 and
around 40% of them are diagnosed with AF
within 12 months.8 Rates of HF were 33% in
paroxysmal, 44% in persistent and 56% in
permanent AF.9 Therefore, the combination
of these two conditions has a dramatic impact
on healthcare and warrants consideration of
new models of care.
HF and AF are closely correlated in terms
of pathophysiology and risk factors.9 10 Owing
to the complex interaction resulting in the
impaired systolic and diastolic function
absent in sinus rhythm, AF can be a cause
or an outcome of HF.11 AF is associated with
significantly increased risk (ie, three times)
of de novo HF.11 On the contrary, develop-
ment and progression of AF are much more
likely to ensue in the presence of structural
and neurohormonal variations seen in HF.12
Moreover, patients with comorbid HF and AF
Protected by copyright. on September 6, 2019 at Sri Lanka:BMJ-PG Sponsored.http://bmjopen.bmj.com/BMJ Open: first published as 10.1136/bmjopen-2019-031033 on 5 September 2019. Downloaded from

2GaoL, MoodieM. BMJ Open 2019;9:e031033. doi:10.1136/bmjopen-2019-031033
Open access
Figure 1 Markov model used for the economic evaluation.
have significantly poorer prognosis irrespective of which
onsets first.13 14 In view of the poor clinical outcome
relating to these two conditions, it is always critical and
challenging to discover the most effective treatment, for
example, treatments effectiveness shown in one condi-
tion or the other independently can be inconsistent with
that revealed in the combined conditions and even raise
safety issues.15 16 This is the case for catheter ablation.
Catheter ablation is a well-established option for symp-
tomatic AF that is resistant to drug therapy in patients
with otherwise normal cardiac function.17–19 Lack of
both clinical evidence and consensus from guidelines20 21
regarding the best management approach for patients
with HF and AF concomitantly leaves a huge knowledge
gap in this field. Very recently, the effectiveness of cath-
eter ablation in improving hard primary endpoints such
as death or the progression of HF in patients comorbid
with HF was tested in a large, randomised controlled
trial.22 The study showed that, after a median follow-up of
37.8 months, the primary composite end point consisting
of death from any cause or worsening of HF that led to
an unplanned hospitalisation, occurred in significantly
fewer patients in the ablation group than the medical
therapy group (HR 0.62, 95% CI 0.43 to 0.87, p=0.007).22
The unanswered question now is whether it is cost-ef-
fective to offer catheter ablation to patients comorbid
with HF and AF given (1) the cost-effectiveness credential
for catheter ablation in AF is not directly applicable to a
patient group with concomitant HF and AF; (2) scarce
healthcare resources.
The primary aim of this study was to assess the lifetime
cost-effectiveness of catheter ablation compared with
conservative medical therapy in treatment patients with
concomitant AF and HF from an Australian healthcare
system perspective.
METHODS
A modelled economic evaluation was performed to assess
the cost-effectiveness of catheter ablation in treating
Australian patients with concomitant AF and HF from
a healthcare system perspective over a lifetime horizon.
More specifically, the patients modelled had a median age
of 64 years, were predominantly male (over 84%), failed
to respond/contraindicated to antiarrhythmic medica-
tions and had one of three types of AF (ie, paroxysmal,
persistent, long-standing persistent). A majority (over
58%) of the modelled population had class II heart func-
tion as rated by the New York Heart Association.22
Model structure
A Markov model was developed to estimate the costs and
health outcomes associated with catheter ablation and
medical therapy for a hypothetical cohort of Australian
patients. The model took a lifetime horizon and the
economic perspective of the model was the Australian
healthcare system. Two health states were considered
(1) alive or (2) dead. The Markov model used a monthly
cycle length with half-cycle correction and assigned each
patient a monthly probability of death based on the time
elapsed and type of treatment received. In each cycle,
the patients who were alive were exposed to the risk of
rehospitalisation due to worsening of HF (readmission to
a hospital for HF-related complications or other causes).
Each patient then accrued lifetime healthcare costs
including treatment (catheter ablation, medications),
outpatient care and examinations, quality-adjusted life
years (QALYs) and life years (LYs) according to their
hospitalisation and treatment status. The model structure
is shown in figure 1. The Markov model was built using
TreeAge (TreeAge Pro 2017, R2.1. TreeAge Software,
Williamstown, Massachusetts, USA).
Model inputs
Transition probabilities
The clinical effectiveness of catheter ablation was derived
from the key clinical study.22 Since the median follow-up
was only 37.8 months, the outcome observed during the
trial was extrapolated beyond the duration of the trial
follow-up. Specifically, the method described by Guyot
et al23 was adopted to derive the individual patient data
(IPD) base on published Kaplan-Meier curves (a vali-
dated graphical digitiser, WebPlotDigitizer V3.9 (http://
arohatgi. info/ WebPlotDigitizer), was used to extract the
graphic data). Parametric survival curves, including expo-
nential, Weibull, log-logistic, log-normal, gompertz and
generalised gamma distributions were fitted to the recon-
structed IPD to extrapolate to a longer time horizon (‘flex-
surv’ package of R). The process of fitting parametric
survival curves to IPD was based on guidance provided
by the Decision Support Unit at the National Institute
for Health and Care Excellence (NICE).24 25 In brief,
Protected by copyright. on September 6, 2019 at Sri Lanka:BMJ-PG Sponsored.http://bmjopen.bmj.com/BMJ Open: first published as 10.1136/bmjopen-2019-031033 on 5 September 2019. Downloaded from

3
GaoL, MoodieM. BMJ Open 2019;9:e031033. doi:10.1136/bmjopen-2019-031033
Open access
this entailed (1) testing of proportional effects assump-
tion (ie, the log cumulative hazard) to determine if the
survival curves of catheter ablation and medical therapy
groups were parallel; (2) then fitting the reconstructed
IPD with single or separate distribution(s) depending on
the conclusion of the first step and (3) determining of the
most appropriate fit by both visual inspection and Akaike
Information Criterion (AIC) and Bayesian Information
Criterion (BIC) goodness-of-fit statistics. If the fitted
curves were similar by distributions (ie, Weibull, expo-
nential and so on) tested in terms of visual inspection,
the most appropriate model was selected based on the
lowest AIC and BIC values. The Kaplan-Meier curves built
on reconstructed IPD are presented in the online supple-
mentary figures 1 and 2. As the proportional hazard
assumption did not hold for the overall survival and time
to event (ie, hospitalisation for worsening HF) curves,
independent parametric curves were fitted separately for
two treatment groups (log cumulative hazard and cumu-
lative hazard for two arms are shown in the online supple-
mentary figures 3 to 4). Based on the criteria set above,
log-normal distribution was selected for the time to hospi-
talisation (see online supplementary figures 5 to 6) of
both catheter ablation and medical therapy arms, while
Weibull and exponential distributions were chosen to
extrapolate the overall survival data for medical therapy
and catheter ablation groups, respectively (see online
supplementary figures 7 to 8). Parameters of the each of
the distributions used to parameterise the fitted curves
are shown in the online supplementary tables 1 and 2.
The time-dependent transition probabilities from alive
to death and from alive to hospitalisation for the first 48
months of the time horizon were directly read from the
published Kaplan-Meier curves.26 From that time-point
beyond, time-dependent transition probabilities were
calculated from the extrapolated curves as described
above. Since the literature indicated that the effectiveness
of AF catheter ablation is likely to be sustained for 3–5
years,27 the aforementioned transition probabilities for
the catheter ablation group were assumed to be the same
as the medical therapy group after 3 years. All the transi-
tion probabilities are presented in the online supplemen-
tary table 3.
Costs
All costs and resource use inputs were obtained from
publicly available sources. The costs taken into account
included: catheter ablation (including hospitalisation
to perform the procedure), medication, examination,
hospitalisation due to worsening of HR, cost related to
death event and adverse events (AEs) related to the cath-
eter ablation procedure. All the costs and resource uses
are presented in tables 1 and 2 and online supplementary
table 4.
Utilities
The baseline utility for patients with HF and the disutility
caused by hospital readmission for HF were incorporated
into the model. The disutility associated with hospitalisa-
tion and/or AEs due to undergoing a catheter ablation
procedure was assumed to be same as the disutility of a
hospital admission for HF and was assumed to be sustained
for 1 year. The sources and utility/disutility values popu-
lated in the model are shown in online supplementary
table 5.
Model assumptions, time horizon, cycle length and perspective
Australian patients who were unresponsive to antiar-
rhythmic medications and diagnosed with both AF and
HF were simulated in the Markov model. The age of the
population was defined as consistent with those recruited
in the pivotal trial. A key assumption of the model was
that the effectiveness of catheter ablation would be main-
tained to 5 years given the median follow-up of the trial
was 37.6 months. The base case time horizon was set to 30
years to capture the lifetime treatment benefit from cath-
eter ablation. However, varied time horizons were exam-
ined in the sensitivity analysis. As Australia has universal
coverage of publicly funded health insurance (ie, Medi-
care), a healthcare system perspective was adopted
to gauge the cost associated with catheter ablation in
patients with AF and HF; a 3% discount rate was applied
for costs, quality-adjusted life years (QALYs) and life years
(LYs). A monthly cycle length with half-cycle correction
was employed to model the risk of events patients may
experience in the long-term.
Cost utility analyses
Incremental cost-effectiveness ratios (ICERs) were calcu-
lated on the basis of two outcomes: QALY and LY gained.
The commonly quoted willingness to pay (WTP) per
QALY threshold of $A50 000 in Australia28 was adopted
to assess whether the catheter ablation was cost-effective.
A cost-effectiveness acceptability curve was constructed to
examine the probability of the intervention being cost-ef-
fective under various WTP/QALY thresholds.
Sensitivity analyses
A series of one-way deterministic sensitivity analyses were
conducted to test the robustness of base case results.
Where applicable, the key model parameters (ie, discount
rate, time horizon, cost of catheter ablation, etc) were
varied within a plausible range informed by literature
or assumptions (see online supplementary table 6). The
results from the one-way sensitivity analyses are shown in
terms of Tornado diagrams, which sequentially graph the
variables with the largest impact on the cost-utility results.
Probabilistic sensitivity analyses (PSA) were performed
to assess the overall impact of uncertainty in the model
by defining distributions for the key parameters (ie, vari-
ables regarding utility and costs) (table 3). Five thousand
iterations were run to construct a mean and 95% CI for
the corresponding costs, and benefits and the results
were plotted on the cost-effectiveness plane.
Patient and public involvement
No patients or public were involved in the study.
Protected by copyright. on September 6, 2019 at Sri Lanka:BMJ-PG Sponsored.http://bmjopen.bmj.com/BMJ Open: first published as 10.1136/bmjopen-2019-031033 on 5 September 2019. Downloaded from

4GaoL, MoodieM. BMJ Open 2019;9:e031033. doi:10.1136/bmjopen-2019-031033
Open access
Table 1 Unit cost of resource use items
Resource uses
% of Patients
using this
resource Source Unit cost Sources of unit cost
Medications*
Antiarrythmic agents 43.5 Roy et al 200839 $A24.01 PBS 2923W, 2876J, 1088G, 2343H, 2344J,
2043M, 8396B
β-blockers 79 Roy et al 200839 $A30.90 PBS 2961W, 3062E, 2565B, 2566B, 2566C,
1081X, 2243C, 8640W, 8605X, 8606Y, 6732N,
8733P, 8743Q, 8735R, 1324Q, 1325R, 9311C,
9312D, 9316H
Long-acting nitrates 17 Roy et al 200839 $A24.39 PBS 11 027J, 11 051P, 1459T, 1515R, 1516T,
3475X, 5108W, 8010N, 8011P, 8026K, 8027 L,
8028M, 8119H, 8171C, 2588F, 1558B, 8273K
Calcium channel blockers 2.5 Roy et al 200839 $A16.19 PBS 2751T, 2752W, 2361G, 2366M, 2367N,
8534E, 8679T, 1694E, 1695F, 1906H, 1907J,
8610E, 1241H, 1248Q, 1250T, 2208F
Digoxin 64.5 Roy et al 200839 $A23.56 PBS 1322N, 2605D, 3164M
ACE-I 86 Roy et al 200839 $A17.43 PBS 1147J, 1148K, 1149 L, 8760C, 1368B,
1369B, 1370D, 1182F, 1183G, 2456G, 2457H,
2458J, 3050M, 3051N,8704D, 9006B, 9007C,
9008D, 1968N, 1969P, 1316G, 1944H, 1945J,
1946K, 8470T, 9120B, 9122D, 2791X, 2793B,
8758Y
ARB 11 Roy et al 200839 $A19.39 PBS 8295N, 8296P, 8297Q, 8889W, 5491B,
8397Y, 8447N, 8951D, 8246B, 8247C, 8248D,
5452Y, 8203R, 2147B, 2148C, 8355R, 8356T,
9368C, 9369D, 9370E, 9371F
Diuretics 44.5 Roy et al 200839 $A36.36 PBS 1484D, 1585K, 2436F, 8532C, 1810G,
1810G, 2411X, 2412Y, 2413B, 2414C, 2415D,
3466K, 8879H, 8880J, 2339D, 2340E
Antiplatelet agents 38.5 Roy et al 200839 $A15.46 PBS 4077N, 10 169F, 2275R, 4179Y, 5436D,
8358X, 9317J, 9354H
Oral anticoagulants† 88 Roy et al 200839 $A69.74 PBS 5054B
Lipid-lowering drug† 43 Roy et al 200839 $A69.72 PBS 10377E
Outpatient care and examinations
Rehabilitation 13.3 Neumanm et al
201540
$A62.25 MBS 10960
Emergency visit‡ 1.2 Neumanm et al
201540
$A1985.00 AR-DRG F62C
GP visits 22.3 Neumanm et al
201540
$A37.05 MBS 23
Specialist visits 5.8 Neumanm et al
201540
$A85.55 MBS 104
Serum urea 16.7 NICE HTA report41 $A9.70 MBS 66500
Electrolytes test 16.7 NICE HTA report41 $A9.70 MBS 66500
Creatinine test 16.7 NICE HTA report41 $A9.70 MBS 66500
GFR§ 16.7 NICE HTA report41 $A9.70 MBS 66500
Hospitalisation care
HF Cost weight $9254.65
With severe complications 2.39 $12 423 AR-DRG F62A
Without severe
complications
1.07 $5548 AR-DRG F62B
Same-day admission 0.58 $3037 AR-DRG F62C
Death due to all causes Per death $5199 Average cost across all AR-DRG items
Continued
Protected by copyright. on September 6, 2019 at Sri Lanka:BMJ-PG Sponsored.http://bmjopen.bmj.com/BMJ Open: first published as 10.1136/bmjopen-2019-031033 on 5 September 2019. Downloaded from

5
GaoL, MoodieM. BMJ Open 2019;9:e031033. doi:10.1136/bmjopen-2019-031033
Open access
Resource uses
% of Patients
using this
resource Source Unit cost Sources of unit cost
*It was assumed that after catheter ablation procedure, patients do not need antiarrhythmic medications. The remaining medications are
the same for both arms.
†According to the Australian Statistics on Medicines 2015, apixaban and atorvastatin+ezetimibe were the mostly prescribed agents.
‡Calculated as the cost/average length of stay=$A3037/1.53 for F62C (HF and shock, transfer less than 5 days).
§The estimated GFR is calculated by the pathology laboratory using the patient’s age, sex and serum creatinine results. Generally
calculated using CKD-EPI (Chronic Kidney Disease Epidemiology Collaboration) formula (https://www.rcpa.edu.au/Library/Practising-
Pathology/RCPA-Manual/Items/Pathology-Tests/E/eGFR).
ACE-I, ACE inhibitor; ARB, angiotensin II receptor blocker; AR-DRG, Australia Adjusted Disease Related Group; GFR, glomerular
ltration rate; GP, general practitioner;HF, heart failure; MBS, medical benets scheme;NICE, National Institute for Health and Care
Excellence; PBS, pharmaceutical benets scheme.
Table 1 Continued
Table 2 Cost of treating catheter ablation-related adverse
events
Adverse events Unit cost Sources
Proportion
(%)
Pericardial effusion $A11
601.00
AR-DRG F61B 1.68
Severe bleeding $A9469.00 AR-DRG
Q62A
1.68
Minor bleeding $A2476.00 AR-DRG
Q62B
1.12
Pulmonary vein
stenosis
$A11
194.00
AR-DRG F10B 0.56
Pneumonia $A5039.00 AR-DRG D63A 1.68
Groin Infection $A5039.00 AR-DRG D63A 0.56
Fever $A5039.00 AR-DRG D63A 0.56
Worsen heart failure $A9254.65 AR-DRG
F62A-C
0.56
The incidence of catheter ablation-related adverse events
was sourced from the study by Marrouche et al 201822, online
supplementary table S11.
AR-DRG, Australian Rened Diagnosis Related Group.
RESULTS
Cost utility analysis
Catheter ablation was associated with higher costs and
benefits (ie, QALYs and LYs) over the lifetime of the
cohort compared with medical therapy alone. The total
cost was $A44 377 per catheter ablation patient and $A28
506 for the medically treated patient, representing an
incremental difference of $A15 871. The primary cost
components in both treatment groups were hospitalisa-
tion ($A6564 in the catheter ablation vs $A5724 in the
medical therapy) and medications ($A14 656 in the cath-
eter ablation vs $A14 534 in the medical therapy) followed
by the outpatient consultations ($A3783 in the catheter
ablation vs $A3539 in the medical therapy). The costs of
AEs associated with the catheter ablation procedure were
$A636 and $A14 063 for the initial and a repeat ($A2977)
of the procedure.
The corresponding QALYs and LYs were 4.58 and 6.99
in the catheter ablation arm, and 4.30 and 6.53 in the
medical therapy arm, which resulted in ICERs of $A55
942/QALY and $A35 020/LY, respectively (table 4).
Based on the normally quoted WTP/QALY threshold
in Australia, offering catheter ablation to patients with
concomitant AF and HF who are not responsive to antiar-
rhythmic medications is not cost-effective.
Sensitivity analyses
The Tornado diagram shows that the ICER was mostly
sensitive to the cost of ablation, time horizon and cost
of outpatient care. The ICER was less sensitive to the
probability of having repeated ablation, baseline utility,
discount rate and cost of death. On contrary, ICER was
not sensitive to the cost of hospitalisation due to wors-
ening of HF. With the variation of these model parame-
ters, the ICER varied to a certain extent (figure 2).
The PSA analyses by incorporating distribution of key
model parameters showed that the mean results based on
5000 simulations of the probabilistic model were identical
to the base case results (table 3). The probability of cath-
eter ablation being not cost-effective was 84% based on
the PSA analysis (figure 3). The cost-effectiveness accept-
ability curve showed that if the WTP/QALY threshold was
greater than $A65 000, catheter ablation may become a
cost-effective treatment strategy in comparison to medical
treatment alone, with a probability of 92.7% (figure 4).
DISCUSSION
The inconsistency in the effectiveness of anti-AF treatment
in the patients with concomitant HF or vice versa is well
recognised. For example, beta blockers are indicated in
patients with symptomatic HF with reduced ejection frac-
tion while their poorer efficacy in patients with concomitant
HF and AF precluded them being used preferentially over
other rate-control medications and not regarded as stan-
dard therapy to improve patients’ prognosis.15 Similarly, it
was observed that adding antiarrythymic drugs for patients
with severe HF led to increased early mortality related to the
worsening of HF.16 It was recognised that incident AF has a
profoundly negative effect on mortality and hospitalisations
for HF with reduced ejection fraction, and it would certainly
appear that the optional time for intervention in patients
Protected by copyright. on September 6, 2019 at Sri Lanka:BMJ-PG Sponsored.http://bmjopen.bmj.com/BMJ Open: first published as 10.1136/bmjopen-2019-031033 on 5 September 2019. Downloaded from

6GaoL, MoodieM. BMJ Open 2019;9:e031033. doi:10.1136/bmjopen-2019-031033
Open access
Table 3 Variables tested and the results from probabilistic sensitivity analyses
Variable Distribution Reference
Cost of hospitalisation due to worsening
of HF
Gamma (alpha 100, lambda 0.0108) Assumption
Cost of death event Gamma (alpha 100, lambda 0.0192) Assumption
Disutility of a hospitalisation due to
worsening of HF
Beta (alpha 89.9, beta 809.1) Yao et al42
Utility of being in HF Beta (alpha 657.72, beta 323.95) Miller et al43
Catheter ablation Medical therapy ICER
Cost $A44 378 (42 628, 46 193) $A28 521 (27 434, 29 705) –
QALYs 4.57 (3.63, 5.43) 4.28 (3.39, 5.09) $A55 234
HF, heart failure; ICER, incremental cost-effectiveness ratio; LYs, life years; QALYs, quality-adjusted life years.
Table 4 Base case results from the Markov model
Catheter
ablation
Medical
therapy ICER
Total cost $A4 377 $A28 506 —
Medication $A14 656 $A14 534 —
Hospitalisation due
to HF
$A6564 $A5724 —
CA and repeated CA* $A14 063 0 —
Examinations $A541 $A506 —
Outpatient
consultation
$A3783 $A3539 —
SAEs $A636 0 —
All cause deaths $A4135 $A4204 —
Number of death† 9991 9992 —
Number of
hospitalisation†
8052 7068 —
QALYs 4.581 4.297 $A55 942/
QALY
LYs 6.985 6.532 $A35 020/LY
*The cost associated with SAEs due to the repeated CA was
included.
†This is based on 10 000 patients.
CA, catheter ablation; HF, heart failure; ICER, incremental cost-
effectiveness ratio; LYs, life years; QALYs, quality-adjusted life
years; SAEs, serious adverse events. Figure 2 Results from the one-way deterministic sensitivity
analysis_ Tornado diagram.
with HF is early after AF onset.29 This triggered exploration
of the effectiveness of catheter ablation in patients with
both AF and HF given its proven effectiveness in patients
with AF. This answered an important clinical question as
to whether to offer this expensive and invasive procedure
to patients deemed at risk of high morbidity and mortality.
The current study however, addressed another unanswered
question in regard to the long-term cost-effectiveness of
catheter ablation for patients comorbid with AF and HF. It
was found that catheter ablation was associated with higher
cost and benefits (ie, QALYs and LYs gained), and the resul-
tant ICER was $A55 942/QALY, which is considered not
cost-effective in the Australia healthcare setting.
AF is one of the most common sustained arrhythmias in
chronic HF. The prognostic influence of the presence of AF
in HF is recognised, with studies reporting an independent
adverse effect on mortality. AF is correlated with left ventric-
ular (LV) systolic function and is associated with an adverse
prognosis in HF regardless of the LV systolic function.14
Hence, treating the AF condition for patients with HF is of
significant importance to improve their long-term survival.
HF is associated with high reoccurring readmission rates
within 30 days of discharge, and a high number of deaths,
nearly half of whom will die within 1 year of discharge.
It was reported that 70% of HF healthcare costs are
attributable to acute hospital care with more than 41 000
people hospitalised annually in Australia (total number
of hospitalisations due to circulatory conditions in Austra-
lia’s hospitals was 556 638 in 2015–16).30 The pressure
to avoid or reduce hospitalisations for patients with HF
is increasing particularly given the Federal government
plans to penalise hospitals for exceeding the benchmark
for readmission rates. From the pivotal trial,22 it was
observed that hospitalisations due to the worsening of HF
were reduced because of the treatment of patients with
catheter ablation, which translates into cost savings in the
long term. This is important and has policy implications
to both the healthcare provider and the Federal govern-
ment. It was worth noting that the benefits in relation
Protected by copyright. on September 6, 2019 at Sri Lanka:BMJ-PG Sponsored.http://bmjopen.bmj.com/BMJ Open: first published as 10.1136/bmjopen-2019-031033 on 5 September 2019. Downloaded from

7
GaoL, MoodieM. BMJ Open 2019;9:e031033. doi:10.1136/bmjopen-2019-031033
Open access
Figure 3 Cost-effectiveness plane. AUD, Asutralian dollar;
QALY, quality-adjusted life years.
Figure 4 Cost-effectiveness acceptability curve. AUD,
Australian dollar; QALY, quality-adjusted life year; WTP,
willingness to pay.
to the treatment of catheter ablation primarily lied with
the reduced mortality due to the delayed worsening of
heart function. For example, our model showed that per
10 000 patients treated, the number of deaths over the
lifetime horizon was 9991 in the catheter ablation arm
versus 9992 in the medical therapy arm or a saving of 1
death. In contrast, the total number of hospitalisations
over the lifetime of the cohort for worsening of HF per
10 000 patients was 8052 in the catheter ablation group
versus 7068 in the medical treatment group (while the
maximum number of hospitalisations per person was
seven in both groups). The higher number of hospital-
isations in the catheter ablation group was due to the
prolonged overall survival of patients (ie, the difference
in LY was 0.453).
The cost-effectiveness of catheter ablation in the treat-
ment of patients with AF has been well studied. In a study
by Chan et al, comparing to medical therapy, catheter abla-
tion had ICERs ranging from US$28 700/QALY to US$51
800/QALY depending on patient characteristics.31 A white
paper by the Institute of Clinical and Economic Review
examined the cost-effectiveness of AF rhythm control strate-
gies in multiple contexts. Catheter ablation was investigated
as first-line and second-line treatments compared with
rate control as a second-line treatment following failure of
amiodarone. The resultant ICERs varied from US$26 869/
QALY (younger patients with low risk) to US$80 615/QALY
(older patients with high risk) for catheter ablation use as
first line; while the ICERs were between US$37 808/QALY
(younger patients with low risk) and US$96 846/QALY
(older patients with high risk) when catheter ablation was
modelled as a second-line therapy.32 A more recent study
by Aronsson et al reported a baseline ICER of €50 570/
QALY when comparing catheter ablation with amiodarone
as a first-line therapy for a lifetime horizon in four Euro-
pean countries. The ICER was lowest in younger patients
(€3434/QALY for those ≤50 years vs €108 937/QALY
for those >50 years).33 Our study was the first to evaluate
the cost-effectiveness of catheter ablation in patients with
concomitant AF and HF. Since the underlying mortality and
morbidity rates are significantly different in patients with
AF alone and AF and HF concomitantly,34 35 the results are
deemed not directly comparable. However, the total QALY
gains for patients with HF predicted by other economic
modelling studies were between 3.99 and 7.7436–38 over a
lifetime horizon, while the results from the current study
fell well within the range.
The greatest strength of the current study is that the
most recent trial data were used to inform the model
parameters. The widely used algorithm was employed to
reconstruct the IPD from the published Kaplan-Meier
curve. The transition probabilities from alive to death
and experiencing hospitalisation (due to worsening of
HF) were directly derived from the reconstructed IPD. In
addition, extensive sensitivity analyses were conducted to
examine the robustness of the base case results. As normal,
this study comes with some limitations. The reconstructed
IPD is only a maximum approximation of the real data.
In particular, the model did not account for the repeated
catheter ablation. However, the cost related to the repeated
procedure was included, and it is believed that the reduced
benefit attributable to the recurrent AF after the catheter
ablation was captured in the pivotal trial, since the median
follow-up was 37.8 months over the study. There are uncer-
tainties around the extrapolation of Kaplan-Meier curves
observed during the trial to the long term; however, the
predicted QALY gains are very similar to the existing
modelled studies in HF. Further, the ICER produced in
the current study may be subject to changes given there
are several ongoing trails examining the catheter ablation
in the same patient population (RAFT-AF NCT01420393,
EAST-AFNET4 NCT01288352, CABANA NCT00911508).
Lastly, since the pivotal trial informed the efficacy of cath-
eter ablation was conducted between 2008 and 2016,
before the advent of state-of-art HF medical therapy (ie,
sacubitril/valsartan), the incremental benefit from the
catheter ablation might be overestimated to some extent,
which would alter the cost-effectiveness conclusion. But
the presented study could still provide important evidence
for its interim cost-effectiveness.
Protected by copyright. on September 6, 2019 at Sri Lanka:BMJ-PG Sponsored.http://bmjopen.bmj.com/BMJ Open: first published as 10.1136/bmjopen-2019-031033 on 5 September 2019. Downloaded from

8GaoL, MoodieM. BMJ Open 2019;9:e031033. doi:10.1136/bmjopen-2019-031033
Open access
CONCLUSIONS
Offering catheter ablation procedure to patients with
systematic paroxysmal or persistent AF who failed to respond
to antiarrthythmic drugs was associated with higher costs
and greater benefits in terms of QALYs and LYs gained in
comparison to medical therapy alone. However, this inter-
vention is not cost-effective from the Australia healthcare
system perspective over a lifetime horizon given its likely
shorter duration of effectiveness.
Acknowledgements LG is supported by the Alfred Deakin Postdoctoral Research
Fellowship, Deakin University, Australia.
Contributors GL conceived and designed the study and drafted the manuscript.
MM interpreted the results and critically revised the manuscript.
Funding The authors have not declared a specic grant for this research from any
funding agency in the public, commercial or not-for-prot sectors.
Competing interests None declared.
Patient consent for publication Not required.
Provenance and peer review Not commissioned; externally peer reviewed.
Data availability statement All data relevant to the study are included in the
article or uploaded as supplementary information.
Open access This is an open access article distributed in accordance with the
Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which
permits others to distribute, remix, adapt, build upon this work non-commercially,
and license their derivative works on different terms, provided the original work is
properly cited, appropriate credit is given, any changes made indicated, and the use
is non-commercial. See:http:// creativecommons. org/ licenses/ by- nc/ 4. 0/.
REFERENCES
1. Braunschweig F, Cowie MR, Auricchio A. What are the costs of heart
failure? Europace 2011;13(suppl 2):ii13–17.
2. Guha K, McDonagh T. Heart failure epidemiology: European
perspective. Curr Cardiol Rev 2013;9:123–7.
3. Wodchis WP, Bhatia RS, Leblanc K, etal. A review of the cost of
atrial brillation. Value in Health 2012;15:240–8.
4. Chugh SS, Havmoeller R, Narayanan K, etal. Worldwide
epidemiology of atrial brillation: a global burden of disease 2010
study. Circulation 2014;129:837–47.
5. Colilla S, Crow A, Petkun W, etal. Estimates of current and future
incidence and prevalence of atrial brillation in the U.S. adult
population. Am J Cardiol 2013;112:1142–7.
6. Australian Institute of Health and Welfare. Australia’s health series no.
12. Cat. no. AUS 122. Canbarra: AIHW, 2010.
7. Desai AS, Stevenson LW. Rehospitalization for heart failure: predict
or prevent? Circulation 2012;126:501–6.
8. Amin AN, Jhaveri M, Lin J. Hospital readmissions in US
atrial brillation patients: occurrence and costs. Am J Ther
2013;20:143–50.
9. Chiang CE, Naditch-Brule L, Murin J, etal. Distribution and risk
prole of paroxysmal, persistent, and permanent atrial brillation
in routine clinical practice: insight from the real-life global survey
evaluating patients with atrial brillation international registry. Circ
Arrhythm Electrophysiol 2012;5:632–9.
10. van Deursen VM, Urso R, Laroche C, etal. Co-Morbidities in patients
with heart failure: an analysis of the European heart failure pilot
survey. Eur J Heart Fail 2014;16:103–11.
11. Stewart S, Hart CL, Hole DJ, etal. A population-based study of the
long-term risks associated with atrial brillation: 20-year follow-up of
the Renfrew/Paisley study. Am J Med 2002;113:359–64.
12. Maisel WH, Stevenson LW. Atrial brillation in heart failure:
epidemiology, pathophysiology, and rationale for therapy. Am J
Cardiol 2003;91:2–8.
13. Wang TJ, Larson MG, Levy D, etal. Temporal relations of atrial
brillation and congestive heart failure and their joint inuence on
mortality. Circulation 2003;107:2920–5.
14. Mamas MA, Caldwell JC, Chacko S, etal. A meta-analysis of the
prognostic signicance of atrial brillation in chronic heart failure. Eur
J Heart Fail 2009;11:676–83.
15. Kotecha D, Holmes J, Krum H, etal. Efcacy of β blockers in patients
with heart failure plus atrial brillation: an individual-patient data
meta-analysis. Lancet 2014;384:2235–43.
16. Køber L, Torp-Pedersen C, McMurray JJV, etal. Increased mortality
after dronedarone therapy for severe heart failure. N Engl J Med
2008;358:2678–87.
17. Wazni OM, Marrouche NF, Martin DO, etal. Radiofrequency ablation
vs antiarrhythmic drugs as rst-line treatment of symptomatic atrial
brillation. JAMA 2005;293:2634–40.
18. Morillo CA, Verma A, Connolly SJ, etal. Radiofrequency ablation
vs antiarrhythmic drugs as rst-line treatment of paroxysmal atrial
brillation (RAAFT-2). JAMA 2014;311:692–9.
19. Nielsen JC, Mortensen LS, Hansen PS. Radiofrequency ablation
as initial therapy in paroxysmal atrial brillation. New Engl J Med
2013;368:478–9.
20. January CT, Wann LS, Alpert JS, etal. 2014 AHA/ACC/HRS guideline
for the management of patients with atrial brillation: a report of the
American College of Cardiology/American heart association Task
force on practice guidelines and the heart rhythm Society. Circulation
2014;130:e199–267.
21. Fuster V, Rydén LE, Cannom DS, etal. ACC/AHA/ESC 2006
guidelines for the management of patients with atrial brillation-
executive summary: a report of the American College of Cardiology/
American heart association Task force on practice guidelines
and the European Society of cardiology Committee for practice
guidelines (writing Committee to revise the 2001 guidelines for
the management of patients with atrial brillation). Eur Heart J
2006;27:1979–2030.
22. Marrouche NF, Brachmann J, Andresen D, etal. Catheter ablation for
atrial brillation with heart failure. N Engl J Med 2018;378:417–27.
23. Guyot P, Ades AE, Ouwens MJNM, etal. Enhanced secondary
analysis of survival data: reconstructing the data from published
Kaplan-Meier survival curves. BMC Med Res Methodol 2012;12:9.
24. Latimer NR. Survival analysis for economic evaluations
alongside clinical trials--extrapolation with patient-level data:
inconsistencies, limitations, and a practical guide. Med Decis Making
2013;33:743–54.
25. Royston P, Parmar MKB. Flexible parametric proportional-hazards
and proportional-odds models for censored survival data, with
application to prognostic modelling and estimation of treatment
effects. Stat Med 2002;21:2175–97.
26. Pharmaceutical Benets Advisory Committee. Guidelines for
preparing submissions to the pharmaceutical benets Advisory
Committee (PBAC). version 5.0, 2016. Available: https:// pbac. pbs.
gov. au/ [Accessed 21 May 2018].
27. Ganesan AN, Shipp NJ, Brooks AG, etal. Long‐term outcomes of
catheter ablation of atrial brillation: a systematic review and Meta‐
analysis. J Am Heart Assoc 2013;2:e004549.
28. Wang S, Gum D, Merlin T. Comparing the ICERs in medicine
reimbursement submissions to NICE and PBAC—Does the presence
of an explicit threshold affect the ICER proposed? Value in Health
2018.
29. Verma A, Kalman JM, Callans DJ. Treatment of patients with atrial
brillation and heart failure with reduced ejection fraction. Circulation
2017;135:1547–63.
30. Australian Institute of Health and Welfare. Admitted patient care
2015–16: Australian hospital statistics. health services series no.75.
cat. No. HSE 185. Canberra: AIHW, 2017.
31. Chan PS, Vijan S, Morady F, etal. Cost-Effectiveness of
radiofrequency catheter ablation for atrial brillation. J Am Coll
Cardiol 2006;47:2513–20.
32. Ollendorf DA, Silverstein MD, Bobo T. Management options for atrial
brillation. US Institute for Clinical and Economic Review, 2010.
33. Aronsson M, Walfridsson H, Janzon M, etal. The cost-effectiveness
of radiofrequency catheter ablation as rst-line treatment for
paroxysmal atrial brillation: results from a MANTRA-PAF substudy.
Europace 2015;17:48–55.
34. Chamberlain AM, Redeld MM, Alonso A, etal. Atrial brillation
and mortality in heart failure: a community study. Circ Heart Fail
2011;4:740–6.
35. Santhanakrishnan R, Wang N, Larson MG, etal. Atrial brillation
begets heart failure and vice versa: temporal associations and
differences in preserved versus reduced ejection fraction. Circulation
2016;133:484–92.
36. Lee D, Wilson K, Akehurst R, etal. Cost-Effectiveness of eplerenone
in patients with systolic heart failure and mild symptoms. Heart
2014;100:1681–7.
37. Almenar L, Díaz B, Quesada A, etal. Cost-Effectiveness analysis
of cardiac resynchronization therapy in patients with NYHA I and
NYHA II heart failure in Spain. Int J Technol Assess Health Care
2013;29:140–6.
38. Grifths A, Paracha N, Davies A, etal. The cost effectiveness of
ivabradine in the treatment of chronic heart failure from the UK
National health service perspective. Heart 2014;100:1031–6.
Protected by copyright. on September 6, 2019 at Sri Lanka:BMJ-PG Sponsored.http://bmjopen.bmj.com/BMJ Open: first published as 10.1136/bmjopen-2019-031033 on 5 September 2019. Downloaded from

9
GaoL, MoodieM. BMJ Open 2019;9:e031033. doi:10.1136/bmjopen-2019-031033
Open access
39. Roy D, Talajic M, Nattel S, etal. Rhythm control versus rate
control for atrial brillation and heart failure. N Engl J Med
2008;358:2667–77.
40. Neumann A, Mostardt S, Biermann J, etal. Cost-Effectiveness and
cost-utility of a structured collaborative disease management in the
interdisciplinary network for heart failure (INH) study. Clin Res Cardiol
2015;104:304–9.
41. Pandor A, Thokala P, Gomersall T, etal. Home telemonitoring or
structured telephone support programmes after recent discharge
in patients with heart failure: systematic review and economic
evaluation. Health Technol Assess 2013;17:1–208.
42. Yao G, Freemantle N, Flather M, etal. Long-Term cost-effectiveness
analysis of nebivolol compared with standard care in elderly patients
with heart failure: an individual patient-based simulation model.
Pharmacoeconomics 2008;26:879–89.
43. Miller LC, Cox KR. Case management for patients with heart
alure: a quality improvement intervention. J Gerontol Nurs
2005;31:20–8.
Protected by copyright. on September 6, 2019 at Sri Lanka:BMJ-PG Sponsored.http://bmjopen.bmj.com/BMJ Open: first published as 10.1136/bmjopen-2019-031033 on 5 September 2019. Downloaded from