EURO PEAN

SO CIETY O F
Review CARDIOLOGY ®

European Journal of Preventive

Cardiology

Cardiac rehabilitation programme after 0(00) 1–10

! The European Society of
Cardiology 2017
transcatheter aortic valve implantation Reprints and permissions:
sagepub.co.uk/journalsPermissions.nav
versus surgical aortic valve replacement: DOI: 10.1177/2047487316686442
journals.sagepub.com/home/ejpc
Systematic review and meta-analysis

Gustavo S Ribeiro1, Rosangela D Melo1, Luı́s F Deresz1,2,

Pedro Dal Lago1,3, Mauro RN Pontes4,5 and Marlus Karsten1,6

Abstract
Background: Aortic stenosis is a valvular heart disease characterised by fixed obstruction of the left ventricular
outflow. It can be managed by surgical aortic valve replacement (sAVR) or transcatheter aortic valve implantation
(TAVI). This review aimed to describe the evidence supporting a cardiac rehabilitation programme on functional capacity
and quality of life in aortic stenosis patients after sAVR or TAVI.
Methods: The search was conducted on multiple databases from January to March 2016. All studies were eligible that
evaluated the effects of a post-interventional cardiac rehabilitation programme in aortic stenosis patients. The meth-
odological quality was assessed using the PEDro scale. Meta-analysis was performed separately by procedure and
between procedures. The walked distance during the six-minute walk test (6MWD) and Barthel index were evaluated.
The analysis was conducted in Review Manager.
Results: Five studies were included (292 TAVI and 570 sAVR patients). The meta-analysis showed that a cardiac
rehabilitation programme was associated with a significant improvement in 6MWD (0.69 (0.47, 0.91); P < 0.001) and
Barthel index (0.80 (0.29, 1.30); P ¼ 0.002) after TAVI and 6MWD (0.79 (0.43, 1.15); P < 0.001) and Barthel index (0.93
(0.67, 1.18); P < 0.001) after sAVR. In addition, the meta-analysis showed that the cardiac rehabilitation programme
promoted a similar gain in 6MWD (4.28% (12.73, 21.29); P ¼ 0.62) and Barthel index (1.52 points (4.81, 1.76);
P ¼ 0.36) after sAVR or TAVI.
Conclusions: The cardiac rehabilitation programme improved the functional capacity and quality of life in aortic stenosis
patients. Patients who underwent TAVI benefitted with a cardiac rehabilitation programme similar to sAVR patients.

Keywords
Aortic valve stenosis, transcatheter aortic valve implantation, surgical aortic valve replacement, exercise, cardiac
rehabilitation
Received 13 October 2016; accepted 6 December 2016

1
Programa de Pós-graduação em Ciências da Reabilitação, Universidade
Federal de Ciências da Saúde de Porto Alegre (UFCSPA), Brazil
2
Departamento de Educação Fı́sica, Universidade Federal de Juiz de Fora
Introduction (UFJF), Brazil
3
Aortic stenosis (AS) is the most common acquired Departamento de Fisioterapia, UFCSPA, Brazil
4
Departamento de Farmacociências, UFCSPA, Brazil
valvular heart disease in the industrialised countries 5
Unidade de Pesquisa, Hospital São Francisco, Brazil
and affects more than 4% of North American and 6
Departamento de Fisioterapia, Universidade do Estado de Santa
European individuals older than 75 years.1,2 Catarina (UDESC), Brazil
Moreover, AS has been associated with higher morbid-
ity and mortality rates than other diseases involving Corresponding author:
cardiac valves.3 The mean survival in the symptomatic Marlus Karsten, Centro de Ciências da Saúde e do Esporte, Universidade
do Estado de Santa Catarina, Rua Pascoal Simone 358, 88080-350,
phase is 2–3 years.4 The main symptoms related to this Florianópolis, SC, Brazil.
disease are chest pain, weakness, syncope or dyspnoea Email: marlus.karsten@udesc.br
during and/or after performing basic activities of daily Twitter: @marluskarsten
2 European Journal of Preventive Cardiology 0(00)
living, decreasing functional independence and quality
of life.5 Surgical aortic valve replacement (sAVR) is the
procedure of choice for patients with symptomatic AS.
However, a significant proportion of these patients has
a high level of frailty, which increases the risk of com-
plications and perioperative mortality. For these
patients, transcatheter aortic valve implantation
(TAVI) has robust therapeutic potential.6
The PARTNER trial was the first large randomised
study that compared sAVR to TAVI in high-risk
patients (PARTNER cohort A) and showed that the
death rates in both groups were similar (35% vs.
34%; P ¼ 0.31).7 In a second study, the authors evalu-
ated patients who were deemed ineligible for surgery
based on the prohibitive surgical risk (PARTNER
cohort B) and randomly assigned them for TAVI or
medical treatment. The patients in the TAVI group
exhibited a significantly lower mortality rate at 3
years when compared to medical treatment (54% vs.
81%; P < 0.001).8 In addition, it has been observed
that death in the medium and long term in the patients
who underwent TAVI was predominantly from non-
cardiac causes.9
A cardiac rehabilitation programme (CRP) may pro-
mote a positive impact in relevant clinical outcomes
and is strongly recommended for patients after cardiac
surgery, especially in the elderly because of the high
prevalence of comorbidities and frailty.9 Various stu-
dies that used different CRP intervention strategies
investigating the safety and effects of CRP in patients
undergoing sAVR or TAVI have recently been pub-
lished.10–14 However, available evidence on the efficacy
and safety of a CRP after AS correction is very limited.
Therefore, the main purpose of this systematic review
and meta-analysis was to summarise the evidence on
the impact of a CRP on functional capacity, exercise
tolerance and health-related quality of life in AS
patients after intervention either by sAVR or TAVI.
Methods
Study design
A systematic review of studies comparing the effects of
a CRP in individuals with corrected AS undergoing
TAVI or sAVR was performed. This review followed
the recommendations of the preferred reporting items
for systematic reviews and meta-analyses (PRISMA)
statement.15 This project was previously registered on
PROSPERO (CRD42015025038).
Data sources and search strategy
We conducted a systematic review, including papers
published from January 2000 to March 2016, with the
following electronic databases: PubMed, EMBASE,
SCOPUS, Web of Knowledge, PEDro, Cochrane
Library and LILACS/BIREME. Relevant studies
were identified using the following MeSH terms:
aortic valve stenosis, transcatheter aortic valve implant-
ation, aortic valve replacement, exercise, cardiac
rehabilitation, oxygen uptake and electric stimulation
therapy. Bibliographies of the included studies and of
the identified related papers were searched manually.
The search was limited to clinical trials with humans
over 18 years old and was not restricted by language.
The full search strategy can be reviewed in
Supplementary Appendix 1.
Eligibility criteria
All eligible studies evaluating the effects of CRP in AS
patients after TAVI or sAVR were considered. There
was no limitation concerning sample size or study
design, and the eligibility criteria did not enforce com-
parisons with a control group. Studies published
between January 2000 and March 2016 were included,
and no trial was excluded based on the presence of
distinct comorbidities at baseline. A priori, editorials
and reviews were deemed ineligible.
Study selection
The results of each search were exported to EndNote
X7 (Thomson Reuters, USA) and two reviewers (GSR
and RDM) assessed the title and abstract independ-
ently. The full texts of all potentially eligible papers
were downloaded to determine their inclusion or exclu-
sion in accordance with the pre-established criteria.
Disagreements were resolved by discussion with the
third assessor (LFD). If any disagreement remained,
the senior researcher (MK) made the final decision.
The k coefficient was used to measure the level of agree-
ment between the reviewers taking into account the clas-
sification proposed by Landis and Koch.16 The authors
provided a narrative synthesis of the included studies
containing the type of intervention, population charac-
teristics and outcomes achieved.
Data extraction and quality assessment
The data were extracted independently (GSR and
RDM). We sought to extract information on: (a) coun-
try of origin; (b) sample size; (c) baseline characteristics,
including mean age, gender, body mass index, New
York Heart Association class, left ventricular ejection
fraction and EuroSCORE; (d) CRP characteristics
(modality of exercise, times per week, intensity and dur-
ation); and (e) outcomes, featuring measures of func-
tional capacity (e.g. oxygen uptake and/or workload)
Ribeiro et al.
and exercise tolerance (e.g. walked distance and exer-
cise time). Measures of secondary outcomes were also
extracted – variables indicating effects on functional
independence (e.g. Barthel index (BI)) and functional
independence measure (FIM)) and health-related qual-
ity of life (e.g. hospital anxiety and depression scale
(HADS) or the EuroQoL visual analogue scale).
Finally, we also collected reported data on all-cause
and cardiovascular mortality as well as safety out-
comes. After extraction, the data were summarised in
a previously standardised worksheet of Excel for
Windows 2010 (Microsoft Corporation, USA).
The methodological quality of studies was assessed
by MRNP and PD using the PEDro scale. Briefly, on
this scale, the reviewers indicated the presence or
absence of 11 quality measures. Considering that criter-
ion 1 was not utilised to calculate the score, the sum of
the others may have led to a maximum of 10 points.
Trials with a score greater than or equal to 6 points
were classified as ‘high-quality trials’, while those with
a score equal to or lower than 5 points were graded as
‘poor quality’.17
Statistical analysis
The meta-analysis included the studies in which the
description of walked distance in the six-minute walk
test (6MWD) before and after a CRP were available in
order to evaluate the effect size of a CRP after AS cor-
rection, and this was done separately for both TAVI
and sAVR. Another meta-analysis comprised studies in
which description of the BI before and after a CRP was
available. For both analyses, we utilised the standar-
dised mean difference (SMD) before and after a CRP
(so-called standardised mean improvement) as the
treatment effect estimate, with a 95% confidence inter-
val (CI), using both fixed-effects and random-effects
models. A SMD higher than 0.5 was considered a mod-
erate, clinically relevant effect size.18 A SMD higher
than 0.8 was determined to be a large, clinically rele-
vant effect size.18
Afterwards, were performed a meta-analysis of the
studies comparing the magnitude of the effect of a post-
operative CRP in TAVI versus sAVR patients, both in
terms of the 6MWD and BI to establish whether a CRP
was more effective after surgical or transcatheter AS
intervention. In this analysis, we assessed the crude,
non-standardised difference between groups as the
treatment-effect estimate, with 95% CI, using a fixed
or random-effects models based on the heterogeneity
of the studies.
Statistical I2 was calculated to assess the heterogen-
eity of the studies included. An I2 score above or near
50% was considered indicative of heterogeneity. In
order to test for publication bias, for each trial, we
3
plotted the effect estimate by the inverse of its standard
error. The symmetry of such ‘funnel plots’ was assessed
both visually and formally with Egger’s test to detect
possible publication bias. Review Manager version 5.3
(Cochrane Collaboration, DNK) was employed for the
statistical analyses and to produce both the funnel and
forest plots.
Results
Search and selection
The systematic search retrieved 1364 references from all
sources. Of these, 100 studies were from PubMed, 401
from EMBASE, 622 from SCOPUS databases, 155
from Web of Science, three from PEDro, 53 from the
Cochrane Library’s basis, 30 from the LILACS/
BIREME databases and four from additional papers.
After the exclusion of duplicates and initial screening
(reading title and abstracts), 19 potentially relevant stu-
dies were identified. According to the study criteria, five
clinical trials were selected for qualitative analysis, all
of them being published in English. These five studies
had sufficient information for quantitative analysis
(meta-analysis). Figure 1 shows the summary of the
search and selection process.
Study characteristics
The characteristics of the included studies are presented
in Table 1. All participants (TAVI 292, sAVR 570) were
selected consecutively in a non-probabilistic conveni-
ence sample and were allocated to a CRP. In general,
studies showed that patients undergoing TAVI were
older, predominantly women and with a EuroSCORE
that was quite poor. Interestingly, European groups in
partnership with regional hospitals developed all
selected studies. The greatest number of publications
was from Italy,11,13,14 followed by Switzerland10 and
Germany.12
Cardiac rehabilitation programme
The CRPs were performed mostly during the early
phase following the procedure (26.9  18.7 days after
valve repair) and within a relatively short period
(2.75  0.45 weeks). In four studies, the CRP length
was 3 weeks,10,12–14 while one trial reported a shorter
length (2 weeks).11 Training frequency ranged from
four to six times per week. In certain studies, the train-
ing programme was carried out at up to two13 or
three10 sessions per day. Two studies did not address
the intensity prescription,12,14 while three studies
reported the use of the BORG scale to control exercise
intensity.10,11,13 Although there was no difference
4 European Journal of Preventive Cardiology 0(00)

Identification
Records identified through Additional records identified
database searching through other sources
(n = 1364) (n = 4)

Records duplicates removed

(n = 273)
Screening

Records screened 1076 records were

(n = 1095) excluded
Eligibility

19 full-text articles 14 Full-text articles were excluded

assesed for eligibility
• Letter to the editor (n = 1)

• Not AS patients (n = 2)

• CRP not described (n = 4)

5 studies included in
qualitative synthesis • Case reports (n = 3)
Included

• Rehabilitation pre valve repair (n = 2)

• Update studies (n = 2)
4 studies included in
quantitative synthesis

Figure 1. Flowchart of the selection of trials included in this systematic review. AS: aortic stenosis; CRP: cardiac rehabilitation
programme.

between the cardiac rehabilitation protocols in the
sAVR and TAVI groups (within studies), the employed
interventions differed among studies. Generally, they
included exercises for endurance training (cycle or
treadmill),10–14 resistance training10–12,14 and/or
inspiratory muscle training.10,11,13 The characteristics
of each study are summarised in Table 2.
Quality assessment
The agreement level of the reviewers was 0.88 (95% CI
0.76–0.99). The PEDro scale oscillated from 3 to 5
points (see Supplementary data). Thus, the methodo-
logical quality of those studies in general was poor
(4.0  0.7 points).16 The trials included in the meta-ana-
lysis were scored at three,13 four10,12 and five points.11
A study investigating only TAVI patients had a score of
four points.14 Overall, no study described a random
allocation process beyond the blinding of subjects, ther-
apists and assessors. The methodological qualities of
each study are shown in Supplementary Appendix 2.
Safety of a CRP in AS patients
The data indicated that a CRP was a safe and effective
form of treatment for patients undergoing valve repair,
regardless of the procedure (sAVR or TAVI) and the
treatment of choice (type, frequency and intensity). In
two studies,10,12 all patients were able to complete the
entire CRP, and there was no report of dropouts,
adverse events or major complications. In one
study,11 there was a very low (1.1%) rate of acute
non-fatal complications that required transference to
the surgical centre (one TAVI patient and one sAVR
patient). In another study,13 in spite of the fact that
TAVI patients were older and more functionally
impaired than sAVR patients, the CRP was safe for
both groups with no report of adverse events related
to the CRP; however, TAVI patients, when compared
with sAVR patients, tolerated a lower workload, had a
worse BI score and were frailer at discharge.
Subsequently, at the mid-term follow-up, the patients
from the TAVI group presented higher mortality com-
pared to the sAVR group (19% vs. 6%; P < 0.05). In
one study evaluating the effects of CRP exclusively in
patients undergoing TAVI,14 these patients were very
frail, dependent and functionally impaired, with a high
risk of clinical complications. In fact, one-third of the
patients evolved with clinical complications; only three
patients (5%) required transference to an acute care
facility. Generally, no patient died during the cardiac
rehabilitation period. These data suggest that
Ribeiro et al. 5

12  9 25  14 <0.001
P value
in patients with AS, those submitted to TAVI were

–
older and sicker than those submitted to sAVR, but
for both groups, the postoperative rehabilitation pro-

22  13
TAVI

–
EuroSCORE
grammes were safe.

sAVR: surgical aortic valve replacement; TAVI: transcatheter aortic valve implantation; BMI: body mass index; NYHA: New York Heart Association; LVEF: left ventricular ejection fraction.
P value sAVR

Effect of CRP on exercise tolerance and functional

–

–
57  8 56  11 0.55 independence after TAVI

57  9 0.38
–

–
The data for the 6MWD were available for all five
trials, which reported data for 286 patients before a
TAVI

CRP and 291 patients after a CRP (Figure 2(a)). The

–
–

–
LVEF (%)

mean walked distance before a CRP was 186  86 m,

56  8
P value sAVR

– increasing to 257  111 m following a CRP. In the

pooled analysis, CRP after TAVI was associated with
–

a significant 6MWD standardised mean improvement

1.9  0.8 2.2  0.8 0.003
2.7  0.8 2.9  0.8 0.22

3.0  0.2 2.9  0.3 0.40

(0.69 (0.47, 0.91); P < 0.001), suggesting a moderate,

clinically relevant effect size.18 There was no evidence
1.8  0.4

for significant heterogeneity between studies (I2 ¼ 40%;

–
TAVI
NYHA classes

P ¼ 0.16).
The data for the BI scoring were available for three
P value sAVR

trials, which reported data for 190 patients before a

–

CRP and 195 patients after a CRP (Figure 2(b)).

The mean BI score before a CRP was 77  19, increas-
27.6  6.1 28.2  5.4 0.66

26.0  3.8 24.7  3.7 0.14

58% <0.001 27.2  4.7 26.1  4.4 0.06

–

ing to 90  14 after a CRP. In the pooled analysis, a

CRP subsequent to TAVI was associated with a signifi-
25.2  4.5

cant BI standardised mean improvement (0.80 (0.29,

–
TAVI

1.30); P ¼ 0.002), indicative of a large, clinically rele-

vant effect size.18 We did detect a significant heterogen-
BMI (kg/m2)

eity across studies with this comparison (I2 ¼ 83%;

–

–
P value sAVR TAVI P value sAVR

P ¼ 0.002). However, we felt that all three studies

pointed in the same direction, showing a significant
0.78

0.10

0.07

standardised mean improvement in BI scoring after a

Table 1. Characteristics of studies included in the methodological analysis.

CRP in patients submitted to TAVI. In addition, a

sensitivity analysis using a fixed-effects model demon-
62%

68%

75%

53%
female (%)

strated very similar results that also featured a

Gender

BI standardised mean improvement (0.74 (0.53, 0.95);

0.001 59%

50%

66%

68  11 80  6 <0.001 34%

P < 0.001).
0.06

Effect of CRP on exercise tolerance and functional

78  6 82  5

81  3 83  4

82  4 86  5

83  5

independence after sAVR

TAVI
Age (years)

The data for the 6MWD were available for four trials,
which included post-sAVR patients; they reported data
sAVR TAVI sAVR

for 570 patients before a CRP and 570 patients after a

CRP (Figure 3(a)). The mean walked distance before
34

64

58

76

60

a CRP was 228  94 m, rising to 315  122 m after a

size (n)
Sample

CRP. In the pooled analysis, a CRP after sAVR was

366
78

74

52

associated with a significant 6MWD standardised mean

Switzerland

improvement (0.93 (0.81, 1.05); P < 0.001). There was

Germany
Country

evidence for significant heterogeneity between studies

Italy

Tarro-Genta Italy

Italy

(I2 ¼ 84%; P < 0.001), leading us to perform a

random-effects model meta-analysis. This analysis indi-
cated a significant 6MWD standardised mean improve-
Fauchère

Zanettini
(2014)

(2014)

(2015)

(2015)

(2014)
Russo
(year)

Völler

ment (0.79 (0.43, 1.15); P < 0.001), a large, clinically

Study

relevant effect size.18

6 European Journal of Preventive Cardiology 0(00)

Table 2. Characteristics of cardiac rehabilitation programme and main findings.

Main
Study (year) Modality Frequency Intensity findings

Fauchère (2014) Gymnastic Daily: 2–3/day Low/medium sAVR
6MWD (82.8  65.1 m);

FIM (12.2  10.9)
Aerobic exercise Weekly: 6/week BORG scale TAVI
6MWD (84.2  68.7 m);

FIM (9.9  6.1)
Respiratory workout Length: 3 weeks
Russo (2014) 30 min of IMT followed Daily: 1/day Low/medium sAVR
6MWD (72.3  57.3 m);
cycling on the morning
Barthel index (10.6  13.2)
30 min callisthenic exercise Weekly: 6/week 70% HRM predict TAVI
6MWD (60.4  46.4 m);
in the afternoon
Barthel index (9.3  12.3)
Length: 2 weeks BORG scale
Tarro-Genta 30 min of cycling or treadmill Daily: 2/day Low/medium sAVR
6MWD (98  95 m);
(2015) starting at minimum
Barthel index (20  16)
workload
30 min of respiratory training Length: 3 weeks BORG scale (14) TAVI
6MWD (94  86 m);

Barthel index (18  15)
Völler (2015) Gymnastics in groups or Daily: 1/day Not described sAVR
6MWD (136.9  135.4 m);
single
ExCap (18.5  30.8 Watts);

HADS (–1.2  3.6)
Cycle and outdoor walking Weekly: 6/week TAVI
6MWD (73.8  89.9 m);

ExCap (10.6  30.3 Watts)
Resistance training Length: 3 weeks
Zanettini (2014) Ambulatory training Daily: 1/day Not described TAVI
6MWD (65  92 m);

Barthel index (11.0  15.5);

EQ-VAS (21.0  12.5)
Bed and sitting exercises Weekly: 6/week
Cycle ergometer/treadmill Length: 3 weeks
Calisthenics exercises
sAVR: surgical aortic valve replacement; 6MWD: six-minute walk distance; FIM: functional independence measure; TAVI: transcatheter aortic valve
implantation; IMT: inspiratory muscle training; HRM: maximal heart rate; ExCap: maximal exercise capacity; HADS: hospital anxiety and depression
scale; EQ-VAS: EuroQoL visual analogue scale.

The data for the BI were available for two trials,
which reported data for 132 patients before a CRP
and 132 patients after a CRP (Figure 3(b)). The mean
BI score before a CRP was 80  21, increasing to 95  9
after a CRP. In the pooled analysis, CRP after sAVR
was associated with a significant BI standardised mean
improvement (0.93 (0.67, 1.18); P < 0.001), a large, clin-
ically relevant effect size.18 There was no evidence for
heterogeneity across studies within this comparison
(I2 ¼ 0%; P ¼ 0.72).
Comparison of effects of CRP in TAVI versus
sAVR patients
Both groups exhibited improved exercise tolerance
(6MWD),10–12,14 functional independence (BI11 and
FIM10 and maximal exercise capacity (workload)12
after a CRP, but only the sAVR group showed
improvement with respect to the HADS.12 One study
did not demonstrate any changes in HADS after the
CRP in both groups10 (Table 2).
The data for studies comparing the effect of a CRP
in TAVI versus sAVR patients for the 6MWD were
available for four studies. They altogether reported
data for 570 sAVR patients and 232 TAVI patients
(Figure 4(a)). The mean additional improvement in
6MWD after a CRP in sAVR patients was 42  37%
(97  88 m) and 50  45% (78  73 m) in TAVI
patients, without difference between procedures in the
pooled analysis (4.28% (12.73, 21.29); P ¼ 0.62).
There was evidence for high heterogeneity among stu-
dies (I2 ¼ 81%; P ¼ 0.001).
Despite its greater diagnostic precision, the cardio-
pulmonary exercise test was used in only one study,
after performing valve repair procedures.11 Although
it was not possible to evaluate the cardiovascular,
Ribeiro et al. 7

(a) After CRP Before CRP Std. Mean Difference Std. Mean Difference
Study or Subgroup Mean SD Total Mean SD Total Weight IV, Random, 95% CI IV, Random, 95% CI
Fauchère et al. 2014 232 133 34 147 102 34 14.3% 0.71 [0.22, 1.20]
Russo et al. 2014 273 108 64 241 95 64 22.2% 0.31 [–0.04, 0.66]
Tarro-Genta et al. 2015 168 131 58 74 58 58 19.9% 0.92 [0.54, 1.31]
Völler et al. 2015 336 86 76 262 90 76 23.5% 0.84 [0.50, 1.17]
Zanettini et al. 2014 275 97 59 210 87 54 20.1% 0.70 [0.32, 1.08]

Total (95% CI) 291 286 100.0% 0.69 [0.47, 0.91]

Heterogeneity: Tau2 = 0.02; Chi2 = 6.66, df=4 (P= 0.16); I2 = 40%
–2 –1 0 1 2
Test for overall effect: Z= 6.16 (P < 0.00001)
6MWD worsening 6MWD improvement

(b)
After CRP Before CRP Std. Mean Difference Std. Mean Difference
Study or Subgroup Mean SD Total Mean SD Total Weight IV, Random, 95% CI IV, Random, 95% CI
Russo et al. 2014 90 17 78 81 24 78 34.9% 0.43 [0.11, 0.75]
Tarro-Genta et al. 2015 84 15 58 66 12 58 32.2% 1.32 [0.91, 1.72]
Zanettini et al. 2014 95 10 59 84 21 54 32.9% 0.67 [0.29, 1.05]

Total (95% CI) 195 190 100.0% 0.80 [0.29, 1.30]

2 2 2
Heterogeneity: Tau = 0.17; Chi = 11.63, df=2 (P= 0.003); I = 83%
–2 –1 0 1 2
Test for overall effect: Z= 3.07 (P = 0.002)
BI worsening BI improvement

Figure 2. Forest plot of the standardised mean improvement on (a) six-minute walked distance and (b) Barthel index in post-TAVI
patients after a CRP.
TAVI: transcatheter aortic valve implantation; CRP: cardiac rehabilitation programme; SD: standard deviation; CI: confidence interval.

(a)
After CRP Before CRP Std. Mean Difference Std. Mean Difference
Study or Subgroup Mean SD Total Mean SD Total Weight IV, Random, 95% CI IV, Random, 95% CI
Fauchère et al. 2014 283 117 78 201 96 78 24.5% 0.76 [0.44, 1.09]
Russo et al. 2014 294 101 74 260 98 74 24.5% 0.34 [0.02, 0.66]
Tarro-Genta et al. 2015 249 132 52 151 80 52 22.1% 0.89 [0.49, 1.30]
Völler et al. 2015 437 140 366 301 102 366 28.9% 1.11 [0.95, 1.26]

Total (95% CI) 570 570 100.0% 0.79 [0.43, 1.15]

2 2 2
Heterogeneity: Tau = 0.11; Chi = 18.84, df=3 (P= 0.0003); I = 84%
–2 –1 0 1 2
Test for overall effect: Z= 4.29 (P < 0.0001)
6MWD worsening 6MWD improvement

(b) After CRP Before CRP Std. Mean Difference Std. Mean Difference
Study or Subgroup Mean SD Total Mean SD Total Weight IV, Fixed, 95% CI IV, Fixed, 95% CI
Russo et al. 2014 98 4 80 88 14 80 60.2% 0.97 [0.64, 1.29]
Tarro-Genta et al. 2015 92 14 52 72 29 52 39.8% 0.87 [0.47, 1.27]

Total (95% CI) 132 132 100.0% 0.93 [0.67, 1.18]

2 2
Heterogeneity: Chi = 0.13, df=1 (P= 0.72); I = 0%
–2 –1 0 1 2
Test for overall effect: Z= 7.16 (P < 0.00001)
BI worsening BI improvement

Figure 3. Forest plot of the standardised mean improvement on (a) six-minute walked distance and (b) Barthel index in post-sAVR
patients after a CRP.
sAVR: surgical aortic valve replacement; CRP: cardiac rehabilitation programme; SD: standard deviation; CI: confidence interval.

ventilatory and metabolic responses to CRP, in patients mean additional improvement in BI scoring after a
that completed the test, exercise capacity at the end of CRP in sAVR patients was 15.3  14.6 points, while
the CRP was similar between groups. Besides that, in the TAVI patients this enhancement was
there was direct correlation between 6MWD and peak 13.7  13.6 points. In the pooled analysis, there was
oxygen consumption, both assessed after CRP. no significant difference between the improvement in
The data for studies comparing the effect of CRP in BI scoring for TAVI versus sAVR patients (1.52
TAVI versus sAVR patients on the BI were available points (4.81, 1.76); P ¼ 0.36). Also, there was no evi-
for two studies that reported data for 132 sAVR dence for heterogeneity across studies via this compari-
patients and 136 TAVI patients (Figure 4(b)). The son (I2 ¼ 0%; P ¼ 0.85).
8 European Journal of Preventive Cardiology 0(00)

(a)
After TAVI After sAVR Mean Difference Mean Difference
Study or Subgroup Mean SD Total Mean SD Total Weight IV, Random, 95% CI IV, Random, 95% CI
Fauchère et al. 2014 56 47 34 41 32 78 23.5% 15.00 [–2.32, 32.32]
Russo et al. 2014 25 19 64 28 22 74 29.6% –3.00 [–9.84, 3.84]
Tarro-Genta et al. 2015 90 82 58 53 52 52 18.3% 37.00 [11.60, 62.40]
Völler et al. 2015 28 34 76 46 45 366 28.6% –18.00 [–26.93, –9.07]

Total (95% CI) 232 570 100.0% 4.28 [–12.73, 21.29]

2 2 2
Heterogeneity: Tau = 242.62; Chi = 24.09, df=3 (P<0.0001); I = 88%
Test for overall effect: Z= 0.49 (P = 0.62) –100 –50 0 50 100
Favours sAVR Favours TAVI

(b) After TAVI After sAVR Mean Difference Mean Difference

Study or Subgroup Mean SD Total Mean SD Total Weight IV, Fixed, 95% CI IV, Fixed, 95% CI
Russo et al. 2014 9.3 12.3 78 10.6 13.2 80 68.1% –1.30 [–5.28, 2.68]
Tarro-Genta et al. 2015 18 15 58 20 16 52 31.9% –2.00 [–7.81, 3.81]

Total (95% CI) 136 132 100.0% –1.52 [–4.81, 1.76]

Heterogeneity: Chi2 = 0.04, df=1 (P = 0.85); I2 = 0%
–20 –10 0 10 20
Test for overall effect: Z= 0.91 (P = 0.36)
Favours sAVR Favours TAVI

Figure 4. Forest plot of the mean difference on (a) six-minute walked distance, in relative values (%), and (b) Barthel index, in points,
post-CRP in TAVI versus sAVR patients.
CRP: cardiac rehabilitation programme; TAVI: transcatheter aortic valve implantation; sAVR: surgical aortic valve replacement; SD:
standard deviation; CI: confidence interval.

Discussion Effect of CRP on exercise tolerance

To our knowledge, this is the first systematic review
investigating CRP effects in patients with severe AS
repaired by either sAVR or TAVI. In both groups, a
CRP increased exercise tolerance, functional independ-
ence, health-related quality of life and maximal exercise
capacity. In addition, the meta-analysis results showed
that patients undergoing sAVR or TAVI had similar
benefits in exercise tolerance and functional independ-
ence. It was not possible to assess objectively the pooled
effect estimates for other variables based on the diver-
sity of the outcomes and assessment tools used in the
clinical trials.
CRP protocols
The CRP consisted of activities aimed at improving
aerobic capacity (walking or cycling) and peripheral
or respiratory muscle resistance. The routine was
engaged in daily for 2 or 3 weeks under supervision
of physiotherapists. The authors did not report the
details of the callisthenic and respiratory workouts, hin-
dering their reproduction in clinical practice and/or in
other research. The greatest contribution was from
Russo et al.,11 who described that the aerobic training
started with 10-minute sessions and increased 5 minutes
per session up to 30 minutes. The intensity did not
exceeded 70% of the predicted maximal heart rate,
monitored by a digital telemetry system. In another
two trials, the prescription and control of intensity
were by the Borg scale of rating of perceived exertion
(up to level 14).10,13
Our data showed a similar gain in exercise tolerance
after the sAVR or TAVI procedure, assessed by the
6MWD. In all trials, both groups exhibited a significant
increase in walked distance without a difference between
groups, even considering the higher age, risk scores and
cardiac comorbidities present in the TAVI group. One
study found that TAVI patients walked 15.9% less than
the sAVR group after propensity score adjustment
(P ¼ 0.004).12 The authors attributed this finding to the
short time that patients stayed in the CRP (sAVR
20.6  3.9 days vs. TAVI 19.2  4.6 days; P ¼ 0.013).
Although there was no study that showed the minimum
clinically important difference (MCID) for the 6MWD
in AS patients who underwent the valve repair proced-
ure, both groups had an average increase that was three
times greater than the MCID observed in CAD patients
who participated in a CRP by 8 weeks after percutan-
eous transluminal coronary angioplasty or coronary
stenting.19 Moreover, Völler et al.12 reported similar
improvements in exercise tolerance (26.7% to sAVR,
P < 0.001; 19.8% to TAVI, P ¼ 0.01) as assessed by a
cycle exercise test. As in other studies,11,13 this work
showed the safety and efficacy of an early and intensive
CRP during the in-hospital phase in elderly patients with
AS undergoing TAVI or sAVR.12
Effect of CRP on functional independence
and health-related quality of life
Similar to exercise tolerance, all patients demonstrated
gains in autonomy and mobility as gauged by the BI.
Ribeiro et al.
In applying this scale, it is possible to assess physical
and cognitive disability by a score that ranges between
0 and 100 points, where high scores indicate a greater
grade of independence. The same authors did not
observe any changes in anxiety and depression levels,
by way of the HADS, after a CRP in both groups.
However, Völler et al.12 did note a reduction in anxiety
and depression with the same scale, emphasising its
subjectivity. In addition, Zanettini et al.14 provided
evidence for improvement in the EuroQoL visual ana-
logue scale, representing a variable of health-related
quality of life (54  14 to 75  11, P < 0.001).
Limitations
Few studies were found that addressed this issue, all
conducted in Europe perhaps because of the greater
life span of the local population,20 particularly in the
countries where the studies were conducted. The poor
methodological quality observed in studies such as this
is usually attributed to experimental design, namely
offering treatment to all patients, which prevented the
blinding and randomisation of the sample. In addition,
the scale adopted in this study (PEDro scale) was devel-
oped to rate the methodological quality of randomised
controlled trials, which differ from our sample, being a
primary limitation of our review. Furthermore, two of
the five papers selected had their methodological evalu-
ation compromised for being published as a ‘Letter to
the Editor’, or a short communication and conference
paper. In addition, one researcher did not make use of a
sAVR group to compare the effect of CRP. However,
as a consequence of the few studies published in this
area, they were determined as worthwhile to include in
our analysis. Finally, another relevant limitation of this
analysis are the several instances of missing information
in the studies, including the lack of standardisation and
poorly detailed characteristics of the cardiac rehabilita-
tion protocols as well as clinical and echocardiographic
variables.
It is important that future studies considering
the inclusion of more data regarding the pre and
post-intervention clinical and echocardiographic vari-
ables, surgical risk scores, frailty assessment, as
well as cardiac rehabilitation protocols (e.g. daily or
weekly frequency, exercise intensity, intensity control
method, exercise time and type of training).
Concerning the methodological quality, although it is
difficult to randomly assign patients according to the
type of valve correction procedure, as well as therapists,
we suggest a group matching and the blinding of the
assessors to the outcomes. In addition, a cardiopulmon-
ary exercise test would provide a more precise measures
for functional diagnosis, exercise prescription and
training monitoring.
9
Conclusion
In conclusion, in patients with AS corrected by sAVR
or TAVI, an early CRP improves the exercise tolerance,
functional independence, health-related quality of life
and maximal exercise capacity. In addition, patients
undergoing sAVR or TAVI had similar benefits in exer-
cise tolerance and functional independence. However,
future studies need to improve their quality of report-
ing, particularly in terms of risk of bias, characteristics
of the CRP protocol and control group.
Author contribution
GSR, RDM and MK contributed to the conception and
design of this study. LFD contributed to the design of this
study. GSR and RDM contributed to the acquisition of the
data. GSR, LFD, PD, MRNP and MK contributed to the
analysis and interpretation of the data. GSR, LFD, MRNP
and MK drafted and critically revised the manuscript. RDM
and PD revised the manuscript. All authors gave final
approval and agree to be accountable for all aspects of the
work ensuring integrity and accuracy.
Declaration of conflicting interests
The authors declared no potential conflicts of interest with
respect to the research, authorship, and/or publication of this
article.
Funding
The authors disclosed receipt of the following financial sup-
port for the research, authorship, and/or publication of this
article: Brazilian National Council for Improvement of
Higher Education (CAPES) master degree scholarships.
References
1. Ozkan A, Kapadia S, Tuzcu M, et al. Assessment of left
ventricular function in aortic stenosis. Nat Rev Cardiol
2011; 8: 494–501.
2. Vahanian A, Alfieri O, Andreotti F, et al. Guidelines on
the management of valvular heart disease (version 2012).
Eur Heart J 2012; 33: 2451–2496.
3. Grimard B, Safford R and Burns E. Aortic stenosis: diagno-
sis and treatment. Am Fam Physician 2016; 93: 371–378.
4. Pereira E, Silva G, Caeiro D, et al. What has changed in
surgical treatment of severe aortic stenosis with the advent
of percutaneous intervention? Rev Port Cardiol 2013; 32:
749–756.
5. Carabello B and Paulus W. Aortic stenosis. Lancet 2009;
373: 956–966.
6. Leon M, Smith C, Mack M, et al. Transcatheter aortic-
valve implantation for aortic stenosis in patients who
cannot undergo surgery. N Engl J Med 2010; 363:
1597–1607.
7. Green P, Cohen D, Généreux P, et al. Relation between
six-minute walk test performance and outcomes after
transcatheter aortic valve implantation (from the
PARTNER trial). Am J Cardiol 2013; 112: 700–706.
10
8. Reynolds M, Magnuson E, Wang K, et al. Cost-effective-
ness of transcatheter aortic valve replacement compared
with standard care among inoperable patients with severe
aortic stenosis: results from the placement of aortic trans-
catheter valves (PARTNER) trial (Cohort B). Circulation
2012; 125: 1102–1109.
9. Singh M, Stewart R and White H. Importance of frailty
in patients with cardiovascular disease. Eur Heart J 2014;
35: 1726–1731.
10. Fauchere I, Weber D, Maier W, et al. Rehabilitation after
TAVI compared to surgical aortic valve replacement. Int
J Cardiol 2014; 173: 564–566.
11. Russo N, Compostella L, Tarantini G, et al. Cardiac
rehabilitation after transcatheter versus surgical pros-
thetic valve implantation for aortic stenosis in the elderly.
Eur J Prev Cardiol 2014; 21: 1341–1348.
12. Voller H, Salzwedel A, Nitardy A, et al. Effect of cardiac
rehabilitation on functional and emotional status in
patients after transcatheter aortic-valve implantation.
Eur J Prev Cardiol 2015; 22: 568–574.
13. Tarro-Genta F, Eleuteri E, Bertolin F, et al. Is cardiac
rehabilitation safe and useful in octogenarians after trans-
catheter aortic valve implantation compared to surgical
aortic valve replacement for aortic stenosis? Eur Heart J
2015; 36(Suppl): 515.
14. Zanettini R, Gatto G, Mori I, et al. Cardiac rehabilita-
tion and mid-term follow-up after transcatheter aortic
valve implantation. J Geriatr Cardiol 2014; 11: 279–285.
European Journal of Preventive Cardiology 0(00)
15. Moher D, Hopewell S, Schulz K, et al. CONSORT 2010
explanation and elaboration: updated guidelines for
reporting parallel group randomised trials. Int J Surg
2012; 10: 28–55.
16. Landis J and Koch G. The measurement of observer
agreement for categorical data. Biometrics 1977; 33:
159–174.
17. Foley N, Bhogal S, Teasell R, et al. Estimates of quality
and reliability with the physiotherapy evidence-based
database scale to assess the methodology of randomized
controlled trials of pharmacological and nonpharmacolo-
gical interventions. Phys Ther 2006; 86: 817–824.
18. Faraone SV. Interpreting estimates of treatment effects:
implications for managed care. Pharm Therapeut 2008;
33: 700–711.
19. Gremeaux V, Troisgros O, Benaim S, et al. Determining
the minimal clinically important difference for the six-
minute walk test and the 200-meter fast-walk test
during cardiac rehabilitation program in coronary
artery disease patients after acute coronary syndrome.
Arch Phys Med Rehabil 2011; 92: 611–619.
20. United Nations. World Population Ageing 2015
[Internet]. New York, USA: Department of Economic
and Social Affairs, Population Division; 2015. Report
No.: ST/ESA/SER.A/390. www.un.org/en/development/
desa/population/publications/pdf/ageing/WPA2015_
Report.pdf (accessed 20 September 2016).