Respiratory Medicine 161 (2020) 105834

Contents lists available at ScienceDirect

Respiratory Medicine
journal homepage: http://www.elsevier.com/locate/rmed

Prevalence and prognosis of respiratory muscle weakness in heart failure

patients with preserved ejection fraction
Nobuaki Hamazaki a, *, Kentaro Kamiya b, Ryota Matsuzawa c, Kohei Nozaki a,
Takafumi Ichikawa a, Shinya Tanaka d, Takeshi Nakamura e, Masashi Yamashita e,
Emi Maekawa f, Chiharu Noda f, Minako Yamaoka-Tojo b, Atsuhiko Matsunaga b,
Takashi Masuda b, Junya Ako f
a
Department of Rehabilitation, Kitasato University Hospital, Sagamihara, Japan
b
Department of Rehabilitation, Kitasato University School of Allied Health Sciences, Sagamihara, Japan
c
Department of Physical Therapy, School of Rehabilitation, Hyogo University of Health Sciences, Kobe, Japan
d
Department of Rehabilitation, Nagoya University Hospital, Nagoya, Japan
e
Department of Rehabilitation Sciences, Kitasato University Graduate School of Medical Sciences, Sagamihara, Japan
f
Department of Cardiovascular Medicine, Kitasato University School of Medicine, Sagamihara, Japan

A R T I C L E I N F O A B S T R A C T

Keywords: Background: Although respiratory muscle weakness (RMW) is known to predict prognosis in patients with heart
Heart failure failure with reduced ejection fraction (HFrEF), RMW prevalence and its prognosis in those with preserved
Preserved ejection fraction ejection fraction (HFpEF) remain unknown. We aimed to investigate whether the RMW predicted mortality in
Respiratory muscle
HFpEF patients.
Prognosis
Methods: We conducted a single-centre observational study with consecutive 1023 heart failure patients (445 in
HFrEF and 578 in HFpEF). Maximal inspiratory pressure (PImax) was measured to assess respiratory muscle
strength at hospital discharge, and RMW was defined as PImax <70% of predicted value. Endpoint was all-cause
mortality after hospital discharge, and we examined the influence of RMW on the endpoint.
Results: Over a median follow-up of 1.8 years, 134 patients (13.1%) died; of these 53 (11.9%) were in HFrEF and
81 (14.0%) in HFpEF. RMW was evident in 190 (42.7%) HFrEF and 226 (39.1%) HFpEF patients and was
independently associated with all-cause mortality in both HFrEF (adjusted hazard ratio [HR]: 2.13, 95% con­
fidence interval [CI]: 1.17–3.88) and HFpEF (adjusted HR: 2.85, 95% CI: 1.74–4.67) patients. Adding RMW to
the multivariate logistic regression model significantly increased area under the receiver-operating characteristic
curve (AUC) for all-cause mortality in HFpEF (AUC including RMW: 0.78, not including RMW: 0.74, P ¼ 0.026)
but not in HFrEF (AUC including RMW: 0.84, not including RMW: 0.82, P ¼ 0.132).
Conclusions: RMW was observed in 39% of HFpEF patients, which was independently associated with poor
prognosis. The additive effect of RMW on prognosis was detected only in HFpEF but not in HFrEF.

that can improve prognosis in HFpEF patients remain unclear, although

some pharmacological or nonpharmacological therapies are recognised
1. Introduction
to improve prognosis in HFrEF patients [7,8]. Therefore, to establish the
proper treatment strategy that can improve prognosis in HFpEF patients,
Heart failure with preserved left ventricular ejection fraction
it is crucial to identify clinically meaningful predictors.
(HFpEF), which is highly observed in elderly patients, comprises
Respiratory muscle weakness (RMW) is frequently observed in pa­
approximately 50% of the overall heart failure patients [1,2]. Patients
tients with chronic heart failure [9,10], and several studies have re­
with HFpEF are known to have exercise intolerance and poor prognosis,
ported that reduced respiratory muscle strength is caused by the atrophy
similar to those with reduced ejection fraction (HFrEF) [3,4].
of these muscles and/or decreased actin-myosin cross-bridges, resulting
Conversely, the clinical characteristics and causes of death differ be­
from the activation of neurohumoral factors due to heart failure
tween patients with HFpEF and HFrEF [5,6]. Additionally, strategies

* Corresponding author. Department of Rehabilitation, Kitasato University Hospital, 1-15-1 Kitasato, Minami-ku, Sagamihara, Kanagawa, 252-0375, Japan.
E-mail address: hamanobu0317@gmail.com (N. Hamazaki).

https://doi.org/10.1016/j.rmed.2019.105834
Received 23 September 2019; Received in revised form 16 November 2019; Accepted 18 November 2019
Available online 20 November 2019
0954-6111/© 2019 Elsevier Ltd. This article is made available under the Elsevier license (http://www.elsevier.com/open-access/userlicense/1.0/).
N. Hamazaki et al.
Abbreviations
BMI body mass index
BNP brain natriuretic peptide
eGFR estimated glomerular filtration rate
HFpEF heart failure with preserved ejection fraction
HFrEF heart failure with reduced ejection fraction
LVEF left ventricular ejection fraction
NYHA New York Heart Association
PImax maximal inspiratory pressure
RMW respiratory muscle weakness
[11–13]. Importantly, RMW, estimated using maximal inspiratory
pressure (PImax), is a known predictor of exercise intolerance and
ventilatory inefficiency, leading to decreased quality of life and lower
survival in HFrEF patients [9,10,14,15]. Conversely, a previous study by
Habedank and colleagues has shown that PImax was not a significant
predictor of mortality as it varied according to gender, body mass index,
and cachexia in severe HFrEF patients [16]. Several statements have
recommended the use of PImax, relative to a reference value (%PImax), for
assessing weakness and dysfunction of respiratory muscles [11,17].
Furthermore, although RMW in HFpEF patients has been reported to
decrease exercise tolerance [18], its prevalence and prognostic potential
in these patients remains unclear. Therefore, we conducted an obser­
vational study to clarify the relationships of RMW assessed by %PImax
with mortality in heart failure patients.
This study aimed to investigate whether RMW predicted mortality in
patients with HFrEF or HFpEF.
2. Methods
The study protocol was approved by the Kitasato Institute Clinical
Research Review Board (KMEO B18-075) and was performed according
to the ethical guidelines of the Declaration of Helsinki.
2.1. Study population
This study had a retrospective longitudinal observational design. We
included consecutive patients with HFrEF or HFpEF who were admitted
to the Kitasato University Hospital for heart failure treatment between
May 2009 and December 2017. HFrEF was defined as a left ventricular
ejection fraction (LVEF) < 40% on an echocardiogram, and HFpEF was
diagnosed based on clinical guidelines and LVEF � 50% [7,19]. Patients
who had undergone thoracic surgery within the last three months or had
chronic respiratory diseases were excluded from the study.
2.2. Study protocol
Patients underwent haematological analysis, echocardiograms, and
assessment of pulmonary function and respiratory muscle function at
hospital discharge. The primary endpoint of this study was defined as
all-cause mortality after hospital discharge. We collected the data on all
variables from electronic database.
2.3. Patient characteristics
Information on age, gender, body mass index (BMI), smoking history,
aetiology of heart failure, severity of heart failure based on the New York
Heart Association functional classification (NYHA class), medications,
and comorbidities such as hypertension, diabetes mellitus, dyslipidae­
mia, chronic kidney disease or atrial fibrillation, was obtained from
medical records at study entry. Routine laboratory analysis included
haemoglobin, serum albumin and C-reactive protein, and plasma brain
2
Respiratory Medicine 161 (2020) 105834
natriuretic peptide (BNP). The estimated glomerular filtration rate
(eGFR) was determined based on serum creatinine levels. Additionally,
echocardiographic variables including LVEF, left atrial diameter (LAD),
mitral early diastolic inflow velocity (E), mitral late diastolic inflow
velocity (A), mitral annular early diastolic velocity (e0 ), and deceleration
time of mitral early diastolic inflow (DCT) were measured. The AHEAD
score was used for risk assessment and was calculated by assigning one
point to each of the following factors: A: atrial fibrillation, H: haemo­
globin <13 g/dL for men and 12 g/dL for women, Elderly (age > 70
years), A: abnormal renal parameters (creatinine > 130 μmol/dL) and D:
diabetes mellitus [20].
2.4. Pulmonary function and respiratory muscle function
Pulmonary function was assessed by measuring forced vital capacity
(FVC) and forced expiratory volume in 1 s (FEV1) using a spirometer
(Autospiro AS-507, Minato Medical Science, Osaka, Japan) and calcu­
lated their percentage based on predictive values issued by the Japanese
Respiratory Society [21]. To assess respiratory muscle function,
maximal inspiratory pressure (PImax) was measured using a pressure
transducer (Autospiro AAM-377, Minato Medical Science, Osaka, Japan)
according to the joint statement of the American Thoracic Society and
European Respiratory Society [17]. To measure PImax, patients in a
sitting position were asked to hold a 25–mm diameter mouthpiece in
their mouth and perform a 3-s forced inspiration from the maximal
expiratory level. PImax was determined based on the average value of the
maximum inspiratory pressure over a 1-s period during the 3-s forced
inspiration. PImax was expressed as its absolute value in the present
study, although it has a negative value compared to atmospheric pres­
sure. Respiratory pressure measurement was performed three times, and
the maximum value in PImax was used for analysis. Subsequently, we
calculated percentage PImax (%PImax) based on predictive values that
were estimated using age, gender, height, and body weight [22]. Res­
piratory muscle weakness (RMW) was defined as %PImax < 70% [23,24].
2.5. Endpoint
The primary endpoint of this study was all-cause death identified
through medical chart review. The time period for this event was
calculated as the number of days from the date of the respiratory muscle
strength measurement to event date. We also investigated whether the
cause of death was cardiovascular (CV) death or non-CV death,
including respiratory diseases.
2.6. Statistical analyses
Differences in clinical parameters between patients with and without
RMW were compared using the unpaired Student’s t-test or the Mann-
Whitney U test for continuous variables and by the Chi-square test or
Fisher’s exact test for categorical variables, as appropriate. We exam­
ined the influence of respiratory muscle weakness on survival using the
Kaplan-Meier method with the log-rank test. To identify predictors of
all-cause mortality, we used a multivariate Cox proportional hazard
model that incorporated clinical characteristics and RMW as covariates.
The subgroup analysis of RMW in various subgroups relevant to the
heart failure prognosis was analysed to assess any potential effect
modification of the association between RMW and all-cause mortality.
To assess the additive effect of RMW on the predictive capability of all-
cause death, the area under the receiver-operating characteristic curves
(AUC) of multivariate logistic regression models for all-cause death were
compared between models with and without RMW. The predictive ac­
curacy and model fit of the logistic regression analyses were examined
using Hosmer-Lemeshow statistics. To confirm whether the sample size
was adequate, we calculated the sample power with an alpha value of
0.05, using mortality and RMW rates, hazard ratio, accrual time during
which patients were recruited, and follow-up time. Continuous variables
N. Hamazaki et al. Respiratory Medicine 161 (2020) 105834

are reported as mean � standard deviation or median with interquartile
range and categorical variables are expressed as patient numbers and
their percentages. A two-tailed P value of <0.05 was considered statis­
tically significant. All statistical analyses were performed using SPSS,
ver. 25.0 (IBM, Armonk, NY), JMP, ver. 14.1.0 (SAS Institute, Cary, NC),
and R, ver. 3.1.2 (R Foundation for Statistical Computing, Vienna,
Austria).
3. Results
3.1. Patient characteristics
The potential study population consisted of 2335 consecutive
patients with HFrEF or HFpEF. Patients who had undergone thoracic
surgery within the last three months (n ¼ 527) or had chronic respira­
tory diseases (n ¼ 175) were excluded, along with those who could not
perform a respiratory muscle function test at hospital discharge (n ¼
610). Consequently, 1023 heart failure patients were included for
analysis; of these 445 had HFrEF and 578 had HFpEF.
Table 1 shows demographic and clinical characteristics of both
HFrEF and HFpEF patients. RMW was observed in 190 (42.7%) HFrEF
patients and 226 (39.1%) HFpEF patients. Among HFrEF patients,
compared to those without RMW, those with RMW were significantly
older, and had higher values for AHEAD score, pack-years, and BNP
levels. HFpEF patients with RMW showed significantly lower BMI values
compared to those without RMW. Both HFrEF and HFpEF patients with

Table 1
Demographic and clinical characteristics in HFrEF and HFpEF patients with (%PImax < 70%) and without (%PImax � 70%) RMW.
HFrEF HFpEF

Variable Overall %PImax � 70% %PImax < 70% Overall %PImax � 70% %PImax < 70%

n (%) 445 255 (57.3) 190 (42.7) P value 578 352 (60.9) 226 (39.1) P value

Age, yrs 63.4 � 15.2 61.5 � 14.8 65.9 � 15.5 0.002 71.4 � 10.8 70.9 � 10.6 72.2 � 11.1 0.154
Gender female, n (%) 123 (27.6) 72 (28.2) 51 (26.8) 0.830 244 (42.2) 155 (44.0) 89 (39.4) 0.301
BMI, kg/cm2 22.4 � 4.4 22.5 � 4.4 22.2 � 4.4 0.582 22.5 � 3.7 22.7 � 3.7 22.0 � 3.6 0.029
SBP, mm Hg 116 � 30 116 � 30 116 � 30 0.829 124 � 29 124 � 28 125 � 29 0.581
DBP, mm Hg 70 � 20 70 � 20 69 � 20 0.422 69 � 17 69 � 16 68 � 18 0.745
HR, beats/min 84 � 21 85 � 22 83 � 20 0.464 79 � 20 80 � 20 77 � 19 0.064
NYHA at hospital discharge, n (%) II 342 (72.9) 216 (84.7) 126 (66.3) <0.001 337 (58.3) 223 (63.4) 114 (50.4) 0.003
III 103 (23.1) 39 (15.3) 64 (33.7) 241 (41.7) 129 (36.6) 112 (49.6)
AHEAD score 1.63 � 1.23 1.47 � 1.16 1.84 � 1.30 0.002 2.09 � 1.15 2.02 � 1.14 2.20 � 1.17 0.064
Prior history of HF, n (%) 157 (35.3) 83 (32.5) 74 (38.9) 0.192 157 (27.2) 92 (26.1) 65 (28.8) 0.503
Smoking history, n (%) 282 (65.3) 161 (65.2) 121 (65.4) 1.000 305 (53.8) 186 (54.1) 119 (53.4) 0.931
Pack-years 37.8 � 32.9 34.3 � 29.1 42.2 � 36.7 0.047 37.3 � 31.0 36.4 � 28.4 38.7 � 34.7 0.535
Medical history
IHD, n (%) 191 (42.9) 109 (42.7) 82 (43.2) 1.000 257 (44.5) 156 (44.3) 101 (44.7) 0.932
Hypertension, n (%) 261 (58.7) 146 (57.3) 115 (60.5) 0.497 397 (68.7) 249 (70.7) 148 (65.5) 0.199
Diabetes mellitus, n (%) 179 (40.2) 101 (39.6) 78 (41.1) 0.770 235 (40.7) 148 (42.0) 87 (38.5) 0.435
Dyslipidaemia, n (%) 212 (47.6) 125 (49.0) 87 (45.8) 0.504 257 (44.5) 162 (46.0) 95 (42.0) 0.391
CKD, n (%) 273 (61.3) 153 (60.0) 120 (63.2) 0.555 404 (70.0) 243 (69.2) 161 (71.2) 0.642
Atrial fibrillation, n (%) 98 (22.0) 49 (19.2) 49 (25.8) 0.106 151 (26.1) 88 (25.0) 63 (27.9) 0.440
Medications
ACE-I, n (%) 261 (58.7) 152 (59.6) 109 (57.4) 0.697 191 (33.0) 118 (33.5) 73 (32.3) 0.786
ARB, n (%) 157 (35.3) 89 (34.9) 68 (35.8) 0.920 260 (45.0) 158 (44.9) 102 (45.1) 1.000
Beta-blockers, n (%) 410 (92.1) 233 (91.4) 177 (93.2) 0.594 388 (67.1) 238 (67.6) 150 (66.4) 0.786
Diuretics, n (%) 401 (90.1) 227 (89.0) 174 (91.6) 0.424 364 (63.0) 214 (60.8) 150 (66.4) 0.186
Blood examination
Haemoglobin, g/dL 13.2 � 2.4 13.4 � 2.5 13.0 � 2.3 0.060 11.9 � 2.1 11.9 � 2.0 11.8 � 2.2 0.635
Albumin, g/dL 3.66 � 0.51 3.71 � 0.51 3.59 � 0.49 0.009 3.53 � 0.53 3.57 � 0.49 3.46 � 0.58 0.009
CRP, mg/dL 0.65 � 1.16 0.69 � 1.32 0.61 � 0.90 0.472 0.92 � 1.43 0.97 � 1.49 0.84 � 1.32 0.276
eGFR, mL/min/1.73m2 52.9 � 23.1 53.9 � 20.9 51.6 � 25.8 0.295 49.3 � 22.9 49.5 � 20.1 49.0 � 26.8 0.807
BNP, pg/mL 390.2 [196.7– 348.4 [181.9 478.5 [215.3 0.014 236.5 [114.6–443.6] 226.2 [115.0 254.8 [115.6 0.090
789.5] –720.0] –898.4] –384.1] –509.6]
Echocardiographic measurements
LVEF, % 28.6 � 7.0 28.7 � 6.9 28.4 � 7.2 0.652 61.5 � 8.1 61.1 � 7.7 62.2 � 8.8 0.103
LAD, mm 44.4 � 8.6 43.9 � 8.4 45.1 � 8.9 0.154 43.3 � 10.4 42.9 � 9.5 43.9 � 11.6 0.267
E/A 1.6 � 1.1 1.6 � 1.1 1.7 � 1.1 0.566 1.2 � 1.0 1.2 � 1.1 1.3 � 0.9 0.383
E0 , cm/s 5.8 � 2.7 5.8 � 2.7 5.7 � 2.7 0.754 6.8 � 3.2 6.7 � 3.2 6.8 � 3.2 0.858
E/E0 16.3 � 7.9 15.9 � 7.8 16.9 � 8.0 0.280 14.8 � 7.5 14.7 � 7.9 14.8 � 6.9 0.873
DCT, ms 176.3 � 67.0 179.4 � 69.5 172.2 � 63.4 0.317 214.6 � 78.4 219.4 � 79.3 207.2 � 76.6 0.113
Respiratory function
FVC, L 2.58 � 0.87 2.79 � 0.84 2.28 � 0.83 <0.001 2.18 � 0.81 2.28 � 0.78 2.02 � 0.83 <0.001
FEV1, L 2.02 � 0.77 2.2 � 0.73 1.77 � 0.75 <0.001 1.67 � 0.67 1.77 � 0.64 1.51 � 0.69 <0.001
FEV1/FVC, % 78.0 � 10.1 78.8 � 8.8 76.9 � 11.5 0.052 76.1 � 9.9 77.6 � 8.2 73.9 � 11.7 <0.001
%FVC, % 76.3 � 17.0 81.4 � 15.1 69.4 � 17.2 <0.001 74.4 � 19.0 78.4 � 17.5 68.0 � 19.6 <0.001
%FEV1, % 73.3 � 18.6 78.5 � 15.8 66.4 � 19.7 <0.001 71.6 � 20.6 76.4 � 18.7 63.9 � 21.1 <0.001
PImax, cmH2O 56.8 � 28.6 73.2 � 25.9 34.8 � 15.7 <0.001 50.1 � 25.9 63.4 � 22.9 29.3 � 14.0 <0.001
%PImax, % 76.2 � 31.6 97.5 � 23.0 47.8 � 14.7 <0.001 79.3 � 35.4 101.1 � 25.7 45.3 � 16.5 <0.001
All-cause mortality, n (%) 53 (11.9) 17 (6.7) 36 (18.9) <0.001 81 (14.0) 26 (7.4) 55 (24.3) <0.001

Values are mean � SD or median [interquartile range].

A, mitral late diastolic inflow velocity; ACE-I, angiotensin convertor enzyme inhibitor; ARB, angiotensin II receptor blocker; BMI, body mass index; BNP, brain
natriuretic peptide; CKD, chronic kidney disease; CRP, C-reactive protein; DBP, diastolic blood pressure; DCT, deceleration time of mitral early diastolic inflow; E,
mitral early diastolic inflow velocity; E0 , mitral annular early diastolic velocity; eGFR, estimated glomerular filtration rate; FEV1, forced expiratory volume in 1-s; FVC,
forced vital capacity; HFpEF, heart failure preserved ejection fraction; HFrEF, heart failure with reduced ejection fraction; HR, heart rate; IHD, ischemic heart disease;
LAD, left atrial diameter; LVEF, left ventricular ejection fraction; NYHA, New York Heart Association; RMW, respiratory muscle weakness; PImax, maximal inspiratory
pressure; SBP, systolic blood pressure.

3
N. Hamazaki et al. Respiratory Medicine 161 (2020) 105834

RMW showed significantly lower albumin levels. Moreover, a higher

percentage of patients with RMW were NYHA class III at hospital
discharge, and patients with RMW also showed significantly lower FVC,
%FVC, FEV1, and %FEV1 values than those without RMW.

3.2. Association between RMW and all-cause mortality

During the median follow-up period of 1.8 years, 134 patients died;
of these, 53 had HFrEF and 81 had HFpEF, and the mortality rate was
72.8/1000 person-years. The cumulative all-cause mortality rate was
6.7% and 18.9% in HFrEF patients without and with RMW, respectively,
and 7.4% and 24.3% in HFpEF patients without and with RMW,
respectively. The Kaplan-Meier survival curves for the two groups are
shown in Fig. 1. Patients with RMW had a significantly lower survival
rate than those without RMW in both HFrEF (log-rank: 16.429, P <
0.001) and HFpEF (log-rank: 38.295, P < 0.001) groups.

3.3. Cause of death

Fig. 2. Rates of cardiovascular death and non-cardiovascular death between
The CV deaths occurred in 93 patients (40 in HFrEF and 53 in HFpEF) presence or absence of RMW in HFrEF and HFpEF. White bar, patients without
and non-CV deaths occurred in 41 patients (13 in HFrEF and 28 in RMW; black bar, patients with RMW. CV, cardiovascular; HFpEF, heart failure
HFpEF). There were no statistical differences in rates of CV and non-CV with preserved ejection fraction; HFrEF, heart failure with reduced ejection
death between HFrEF and HFpEF (Supplementary file). Patients with fraction; RMW, respiratory muscle weakness.
RMW showed significantly higher rates of CV death (P ¼ 0.026 for
HFrEF and P ¼ 0.027 for HFpEF) and non-CV death (P ¼ 0.012 for HFrEF
Table 2
and P ¼ 0.005 for HFpEF) compared to those without RMW, both in
Cox proportional hazard models of respiratory muscle weakness for all-cause
HFrEF and HFpEF (Fig. 2).
mortality in HFrEF and HFpEF.
%PImax � 70% %PImax < 70%
3.4. Cox proportional hazard models for RMW and all-cause death
HR 95% CI HR 95% CI P value

Table 2 shows the results of the Cox proportional hazard models for HFrEF
RMW and all-cause mortality. In the univariate Cox proportional hazard Univariate analysis 1 Reference 3.09 1.74–5.48 <0.001
Multivariate analysis 1 Reference 2.13 1.17–3.88 0.011
model, RMW, defined as %PImax < 70%, was a significant predictor of
all-cause mortality in HFrEF and HFpEF patients. The multivariate HFpEF
model identified RMW as a significant independent predictor for all- Univariate analysis 1 Reference 3.93 2.46–6.27 <0.001
cause mortality even after adjustment for clinical confounding factors, Multivariate analysis 1 Reference 2.85 1.74–4.67 <0.001
both in HFrEF and in HFpEF patients. No significant interactions were Multivariate analyses were adjusted for age, gender, BMI, AHEAD score, NYHA
observed in the association between RMW and poor prognosis across the class at hospital discharge, IHD, LVEF, and BNP. CI, confidence interval; HFrEF,
various subgroups in both HFrEF and HFpEF patients, even after heart failure with reduced ejection fraction; HFpEF, heart failure with preserved
adjustment for confounding factors used in the multivariate Cox pro­ ejection fraction; HR, hazard ratio; PImax, maximal inspiratory pressure.
portional hazard model (Fig. 3). The sample size in this study was suf­
ficient, as reflected by a sample power of HFrEF and HFpEF of 0.995 and

Fig. 1. Kaplan-Meier survival curves for the association between respiratory muscle weakness and all-cause death in HFrEF and HFpEF. (A), patients with HFrEF and
(B), patients with HFpEF. Solid line, patients without RMW; dotted line, patients with RMW. HFpEF, heart failure with preserved ejection fraction; HFrEF, heart
failure with reduced ejection fraction; PImax, maximal inspiratory pressure; RMW, respiratory muscle weakness.

4
N. Hamazaki et al. Respiratory Medicine 161 (2020) 105834

0.998, respectively. not in those with HFrEF.

4.1. Previous studies

3.5. Additive effect of RMW on predictive capability for all-cause death

To the best of our knowledge, this is the first study to demonstrate

The AUCs of the multivariate logistic regression models for all-cause
that 39% of HFpEF patients have RMW and that it is a significant indi­
mortality are shown in Fig. 4. The AUC of a model that used clinical
cator of poor prognosis in these patients. Previous studies have reported
characteristics as covariates in HFrEF patients was 0.82 (95% CI:
RMW in 30%–50% of HFrEF patients [11], which is similar to that seen
0.75–0.87), and the addition of RMW to the model did not increase the
in the present study. Thus, we show that the prevalence of RMW is
AUC (AUC: 0.84, 95% CI: 0.78–0.88, P ¼ 0.132). In HFpEF patients, the
comparable among HFpEF and HFrEF patients. Conversely, mean ab­
AUC of the model without RMW was 0.74 (95% CI: 0.68–0.80), which
solute value of PImax was significantly lower in HFpEF patients (50.1
significantly increased to 0.78 (95% CI: 0.72–0.83, P ¼ 0.026) when
cmH2O) than in HFrEF patients (56.8 cmH2O), suggesting that the value
RMW was included in the model. In the Hosmer-Lemeshow statistics of
of PImax as a respiratory muscle strength in heart failure differs between
logistic regression models, both the HFrEF and HFpEF models reached
HFpEF and HFrEF patients (Supplementary file). Generally, respiratory
statistical significance for predicting all-cause mortality (HFrEF: chi-
muscle strength is lower in older patients, females, and in patients with
squared ¼ 7.47, predictive value ¼ 88%, P ¼ 0.487; HFpEF: chi-squared
severe heart failure [16]. Further, our study population tended to
¼ 11.19, predictive value ¼ 86%, P ¼ 0.191).
include older patients and a higher proportion of females in the HFpEF
group compared to the HFrEF group, which is consistent with the re­
4. Discussion
ported higher prevalence of HFpEF in the elderly and in females [2].
Additionally, it is notable that %PImax levels were comparable between
The novel findings of this study are as follows. First, RMW was
HFrEF and HFpEF patients in the present study. Thus, we believe that
observed in 39% of the patients with HFpEF. Second, both HFrEF and
the use of %PImax values would be both useful and important in assessing
HFpEF patients with RMW had a significantly higher mortality rate
RMW for heart failure patients.
compared to those without RMW, and RMW increased the risk of all-
cause mortality in HFrEF and HFpEF patients by two- and three-fold,
respectively. However, the additive effect of RMW on predictive capa­
bility for poor prognosis was observed only in patients with HFpEF but

Fig. 3. Forest plots of hazard ratios for the association between respiratory muscle weakness and all-cause mortality according to major subgroups. Hazard ratios
were adjusted for age, gender, BMI, AHEAD score, NYHA class at hospital discharge, ischemic heart disease, LVEF, and BNP. ACE-I, angiotensin convertor enzyme
inhibitor; ARB, angiotensin II receptor blocker; BMI, body mass index; BNP, brain natriuretic peptide; HFpEF, heart failure with preserved ejection fraction; HFrEF,
heart failure with reduced ejection fraction; LVEF, left ventricular ejection fraction; NYHA, New York Heart Association functional classification.

5
N. Hamazaki et al. Respiratory Medicine 161 (2020) 105834

Fig. 4. Receiver-operating characteristic curves of logistic regression models to predict all-cause mortality in HFrEF and HFpEF. (A), patients with HFrEF and (B),
patients with HFpEF. Independent variables of the logistic regression model (model 1) were age, gender, BMI, AHEAD score, NYHA class at hospital discharge,
ischemic heart disease, LVEF, and BNP. Solid line, model 1 with RMW; Dotted line, model 1 without RMW. AUC, area under the receiver-operating characteristic
curve; BMI, body mass index; BNP, brain natriuretic peptide; CI: confidence interval; HFpEF, heart failure with preserved ejection fraction; HFrEF, heart failure with
reduced ejection fraction; LVEF, left ventricular ejection fraction; NYHA, New York Heart Association functional classification; RMW, respiratory muscle weakness.

4.2. Interpretations of findings
Although there were no differences in age, gender, smoking habits,
comorbidities, or medications between HFpEF patients with and without
RMW, patients with RMW showed significantly lower pulmonary func­
tion, as evidenced by the lower FVC and FEV1 values, and a higher
mortality rate compared to those without RMW. This observation is
partially different from that seen in HFrEF patients, i.e., RMW in HFrEF
was associated with older age, presence of comorbid conditions, smok­
ing habits, and higher BNP. Bowen et al. have demonstrated that acti­
vated levels of ROS and the ubiquitin proteasome system in respiratory
muscles, which are known to cause muscle atrophy in HFrEF [25], were
not elevated in rats with HFpEF even though these animals also had
significant diaphragm muscle atrophy that resulted in RMW [26]. These
molecular and histological alterations could have contributed to the
observed differences between HFrEF and HFpEF patients with or
without RMW. Furthermore, although RMW was identified as an inde­
pendent predictor of all-cause mortality in both HFrEF and HFpEF in this
study, additional effect of RMW to traditional prediction model of heart
failure that included NYHA and comorbid conditions was observed only
in HFpEF patients but not in HFrEF patients. Recent studies on the cause
of death in heart failure patients have reported that while the main
causes of death in HFrEF patients were heart failure exacerbation and
sudden death, non-cardiovascular death, including due to respiratory
failure or infections of the respiratory system, were the main causes of
death in HFpEF patients, apart from cardiovascular causes of death [5,
6]. In general, the decline in respiratory muscle strength is associated
with inefficient ventilation as a cause of dyspnoea [27], and with
reduced pulmonary function [9,15], which is a known risk factor for
heart failure and/or respiratory infection [28–30]. Therefore, in the
present study, RMW worsened HFpEF patient prognosis because it can
decrease pulmonary function that leads to respiratory complications and
the incidence of cardiovascular events, even in the subgroup analysis
stratified into the previously reported indicator of HFpEF.
4.3. Clinical implications
been identified as a new predictor of prognosis in HFpEF patients. As
respiratory muscle strength is easily measured in clinical practice, RMW
might be a useful marker for risk classification not only in HFrEF but also
in HFpEF patients. Furthermore, an increase in respiratory muscle
strength due to inspiratory muscle training has been reported to improve
exercise tolerance and quality of life in HFrEF patients [23,31], and our
results imply similar potential benefits due to greater respiratory muscle
strength in HFpEF patients as well.
4.4. Potential limitations
There are some limitations in the present study. As this was a single-
centre study that only included Japanese patients, it is unclear whether
these results can apply to patients in other hospitals or in other pop­
ulations. Nevertheless, the sample size used here is larger than that of
previous studies on respiratory muscle strength in HFpEF patients [18,
24], which statistically satisfied the power analysis for estimating
sample size. However, given half of potential study population was
excluded from the analysis, external validity could have been reduced.
We also performed the multivariate analyses using multiple con­
founders. Such multiple test might increase the rate of false positives
(type I error). Therefore, future multi-centre prospective studies are
required to investigate the validity and reliability of RMW, assessed by
%PImax, as a predictor of prognosis in these patients.
5. Conclusions
RMW, defined as %PImax < 70%, was independently associated with
poor prognosis in both HFrEF and HFpEF patients. However, the addi­
tive effect of RMW on risk prediction was observed only in HFpEF and
not in HFrEF patients.
Funding
This work was supported by Japan Society for the Promotion of
Science Grant-in-Aid [JSPS KAKENHI Grant Number JP16K16442].

The results presented here have clinical implications that RMW has

6
N. Hamazaki et al.
Author contributions
NH, KK, and TM contributed to the conception and design of the
study. NH and TM wrote the manuscript. NH, KK, RM, KN, TI, ST, TN,
and MY contributed to data collection. NH, TM, KK, RM, KN, ST, EM,
CN, MT, AM, and JA contributed to interpretation. NH and KK
contributed to the statistical analysis. JA contributed to supervision and
mentorship. All authors have critically revised and assisted in the
preparation of the manuscript. All gave final approval and agree to be
accountable for all aspects of work ensuring integrity and accuracy.
Research ethics and patient consent
Ethical approval for the study was given by Kitasato Institute Clinical
Research Review Board (KMEO B18-075).
Declaration of competing interest
None.
Appendix A. Supplementary data
Supplementary data to this article can be found online at https://doi.
org/10.1016/j.rmed.2019.105834.
References
[1] T.E. Owan, D.O. Hodge, R.M. Herges, S.J. Jacobsen, V.L. Roger, M.M. Redfield,
Trends in prevalence and outcome of heart failure with preserved ejection fraction,
N. Engl. J. Med. 355 (2006) 251–259.
[2] C.S. Lam, E. Donal, E. Kraigher-Krainer, R.S. Vasan, Epidemiology and clinical
course of heart failure with preserved ejection fraction, Eur. J. Heart Fail. 13
(2011) 18–28.
[3] M.J. Farr, C.C. Lang, J.J. Lamanca, M.R. Zile, G. Francis, L. Tavazzi, W.H. Gaasch,
M. St John Sutton, H. Itoh, D. Mancini, M.-G. Investigators, Cardiopulmonary
exercise variables in diastolic versus systolic heart failure, Am. J. Cardiol. 102
(2008) 203–206.
[4] B.A. Borlaug, W.J. Paulus, Heart failure with preserved ejection fraction:
pathophysiology, diagnosis, and treatment, Eur. Heart J. 32 (2011) 670–679.
[5] M. Vaduganathan, R.B. Patel, A. Michel, S.J. Shah, M. Senni, M. Gheorghiade,
J. Butler, Mode of death in heart failure with preserved ejection fraction, J. Am.
Coll. Cardiol. 69 (2017) 556–569.
[6] S. Hamaguchi, S. Kinugawa, M.A. Sobirin, D. Goto, M. Tsuchihashi-Makaya,
S. Yamada, H. Yokoshiki, H. Tsutsui, J.-C. Investigators, Mode of death in patients
with heart failure and reduced vs. preserved ejection fraction: report from the
registry of hospitalized heart failure patients, Circ. J. 76 (2012) 1662–1669.
[7] P. Ponikowski, A.A. Voors, S.D. Anker, H. Bueno, J.G. Cleland, A.J. Coats, V. Falk,
J.R. Gonzalez-Juanatey, V.P. Harjola, E.A. Jankowska, et al., ESC Guidelines for the
diagnosis and treatment of acute and chronic heart failure: the Task Force for the
diagnosis and treatment of acute and chronic heart failure of the European Society
of Cardiology (ESC). Developed with the special contribution of the Heart Failure
Association (HFA) of the ESC, Eur. J. Heart Fail. 18 (2016) 891–975, 2016.
[8] M. Senni, W.J. Paulus, A. Gavazzi, A.G. Fraser, J. Diez, S.D. Solomon, O.A. Smiseth,
M. Guazzi, C.S. Lam, A.P. Maggioni, et al., New strategies for heart failure with
preserved ejection fraction: the importance of targeted therapies for heart failure
phenotypes, Eur. Heart J. 35 (2014) 2797–2815.
[9] F.J. Meyer, M.M. Borst, C. Zugck, A. Kirschke, D. Schellberg, W. Kubler, M. Haass,
Respiratory muscle dysfunction in congestive heart failure: clinical correlation and
prognostic significance, Circulation 103 (2001) 2153–2158.
[10] L. Frankenstein, M. Nelles, F.J. Meyer, C. Sigg, D. Schellberg, B.A. Remppis, H.
A. Katus, C. Zugck, Validity, prognostic value and optimal cutoff of respiratory
muscle strength in patients with chronic heart failure changes with beta-blocker
treatment, Eur. J. Cardiovasc. Prev. Rehabil. 16 (2009) 424–429.
Respiratory Medicine 161 (2020) 105834
[11] R.C. Kelley, L.F. Ferreira, Diaphragm abnormalities in heart failure and aging:
mechanisms and integration of cardiovascular and respiratory pathophysiology,
Heart Fail. Rev. 22 (2017) 191–207.
[12] F. Laghi, M.J. Tobin, Disorders of the respiratory muscles, Am. J. Respir. Crit. Care
Med. 168 (2003) 10–48.
[13] G. Stassijns, G. Gayan-Ramirez, P. De Leyn, G. Verhoeven, P. Herijgers, V. de Bock,
R. Dom, R. Lysens, M. Decramer, Systolic ventricular dysfunction causes selective
diaphragm atrophy in rats, Am. J. Respir. Crit. Care Med. 158 (1998) 1963–1967.
[14] T.P. Chua, S.D. Anker, D. Harrington, A.J. Coats, Inspiratory muscle strength is a
determinant of maximum oxygen consumption in chronic heart failure, Br. Heart J.
74 (1995) 381–385.
[15] N. Hamazaki, T. Masuda, K. Kamiya, R. Matsuzawa, K. Nozaki, E. Maekawa,
C. Noda, M. Yamaoka-Tojo, J. Ako, Respiratory muscle weakness increases dead-
space ventilation ratio aggravating ventilation-perfusion mismatch during exercise
in patients with chronic heart failure, Respirology 24 (2019) 154–161.
[16] D. Habedank, F.J. Meyer, R. Hetzer, S.D. Anker, R. Ewert, Relation of respiratory
muscle strength, cachexia and survival in severe chronic heart failure, J Cachexia
Sarcopenia Muscle 4 (2013) 277–285.
[17] American thoracic society/European respiratory S: ATS/ERS statement on
respiratory muscle testing, Am. J. Respir. Crit. Care Med. 166 (2002) 518–624.
[18] K. Yamada, Y. Kinugasa, T. Sota, M. Miyagi, S. Sugihara, M. Kato, K. Yamamoto,
Inspiratory muscle weakness is associated with exercise intolerance in patients
with heart failure with preserved ejection fraction: a preliminary study, J. Card.
Fail. 22 (2016) 38–47.
[19] C.W. Yancy, M. Jessup, B. Bozkurt, J. Butler, D.E. Casey Jr., M.H. Drazner, G.
C. Fonarow, S.A. Geraci, T. Horwich, J.L. Januzzi, et al., ACCF/AHA guideline for
the management of heart failure: a report of the American college of cardiology
foundation/American heart association task force on practice guidelines, J. Am.
Coll. Cardiol. 62 (2013) e147–239, 2013.
[20] Y.J. Chen, S.H. Sung, H.M. Cheng, W.M. Huang, C.L. Wu, C.J. Huang, P.F. Hsu, J.
S. Yeh, C.Y. Guo, W.C. Yu, C.H. Chen, Performance of AHEAD score in an asian
cohort of acute heart failure with either preserved or reduced left ventricular
systolic function, J Am Heart Assoc 6 (2017).
[21] N. Tojo, H. Suga, M. Kambe, [Lung function testing –the official guideline of the
Japanese respiratory society], Rinsho Byori 53 (2005) 77–81.
[22] M. Suzuki, S. Teramoto, E. Sudo, K. Ogawa, T. Namekawa, K. Motrita, T. Matsuse,
H. Takizawa, Y. Ouchi, Y. Fukuchi, [Age-related changes in static maximal
inspiratory and expiratory pressures], Nihon Kyobu Shikkan Gakkai Zasshi 35
(1997) 1305–1311.
[23] P. Dall’Ago, G.R. Chiappa, H. Guths, R. Stein, J.P. Ribeiro, Inspiratory muscle
training in patients with heart failure and inspiratory muscle weakness: a
randomized trial, J. Am. Coll. Cardiol. 47 (2006) 757–763.
[24] P. Palau, E. Dominguez, E. Nunez, J.M. Ramon, L. Lopez, J. Melero, A. Bellver, F.
J. Chorro, V. Bodi, A. Bayes-Genis, et al., Inspiratory muscle function and exercise
capacity in patients with heart failure with preserved ejection fraction, J. Card.
Fail. 23 (2017) 480–484.
[25] H.W. van Hees, H.F. van der Heijden, C.A. Ottenheijm, L.M. Heunks, C.J. Pigmans,
F.W. Verheugt, R.M. Brouwer, P.N. Dekhuijzen, Diaphragm single-fiber weakness
and loss of myosin in congestive heart failure rats, Am. J. Physiol. Heart Circ.
Physiol. 293 (2007) H819–H828.
[26] T.S. Bowen, N.P. Rolim, T. Fischer, F.H. Baekkerud, A. Medeiros, S. Werner,
E. Bronstad, O. Rognmo, N. Mangner, A. Linke, et al., Heart failure with preserved
ejection fraction induces molecular, mitochondrial, histological, and functional
alterations in rat respiratory and limb skeletal muscle, Eur. J. Heart Fail. 17 (2015)
263–272.
[27] J.A. Neder, D.C. Berton, M. Marillier, A.C. Bernard, O.D. DE, N. Canadian
Respiratory Research, The role of evaluating inspiratory constraints and
ventilatory inefficiency in the investigation of dyspnea of unclear etiology, Respir.
Med. 158 (2019) 6–13.
[28] S.K. Agarwal, G. Heiss, R.G. Barr, P.P. Chang, L.R. Loehr, L.E. Chambless,
E. Shahar, D.W. Kitzman, W.D. Rosamond, Airflow obstruction, lung function, and
risk of incident heart failure: the Atherosclerosis Risk in Communities (ARIC)
study, Eur. J. Heart Fail. 14 (2012) 414–422.
[29] R.O. Crapo, Pulmonary-function testing, N. Engl. J. Med. 331 (1994) 25–30.
[30] R. Andrea, A. Lopez-Giraldo, C. Falces, T. Lopez, L. Sanchis, C. Gistau, M. Sabate,
M. Sitges, J. Brugada, A. Agusti, Pulmonary function predicts mortality and
hospitalizations in outpatients with heart failure and preserved ejection fraction,
Respir. Med. 134 (2018) 124–129.
[31] N.A. Smart, F. Giallauria, G. Dieberg, Efficacy of inspiratory muscle training in
chronic heart failure patients: a systematic review and meta-analysis, Int. J.
Cardiol. 167 (2013) 1502–1507.