High prevalence of occult left ventricular diastolic dysfunction detected by exercise stress test in systemic sclerosis

Despite the poor prognosis of systemic sclerosis (SSc) due to the co-occurrence of left ventricular diastolic dysfunction (LVDD), presence of occult LVDD has not been sufficiently investigated. This retrospective study aimed to reveal the prevalence and determinants of occult LVDD in patients with SSc by exercise stress test. Forty-five SSc patients (age, 63 ± 13 years; men/women, 6/39) with normal pulmonary artery pressure and pulmonary artery wedge pressure (PAWP) at rest underwent a symptom-limited exercise test with right heart catheterization using a supine cycle ergometer; haemodynamic parameters at rest, leg raise and during exercise were evaluated. Occult LVDD defined PAWP ≥ 25 mmHg during exercise was seen in 13 patients (29%). Higher PAWP, lower pulmonary vascular resistance and diastolic pulmonary pressure gradient, larger left atrium at rest, and higher PAWP during leg raise (15 ± 4 vs 10 ± 4 mmHg in non-LVDD group, p < 0.001) were observed in the occult LVDD group. The area under the ROC curve for PAWP after leg raise was largest at 0.83 (95% CI: 0.70–0.95, p = 0.001). About one-third (29%) of SSc patients with normal haemodynamics at rest showed occult LVDD. A higher PAWP after leg raise could be useful for detecting occult LVDD.

PAP of 21-24 mmHg, called borderline PH, were recently established, and the importance of early differentiation between pre-capillary and post-capillary PH was emphasized. Therefore, we investigated the prevalence of occult LVDD in patients with SSc who were referred for clinical assessment because of dyspnoea during exertion via the exercise stress test with normal haemodynamics at rest. We also examined the factors predicting PAWP elevation during exercise among the resting haemodynamic parameters at rest and during leg raise.

Results
Baseline characteristics. The patient flow chart is shown in Fig. 1. Among the 50 patients who underwent exercise testing, five were excluded due to a high mean PAP. No ECG changes were observed during exercise. Eventually, 45 patients (age, 63 ± 13 years; men/women, 6/39) were analysed. Of these, 13 (29%) showed occult LVDD. The distributions of the PAWP at rest by group are shown in Fig. 2. The baseline characteristics of the study groups classified by the presence or absence of occult LVDD are shown in Table 1. Parameters determined  www.nature.com/scientificreports/ ties, pulmonary function, mean PAP, CO, and 6-min walk distance (6MWD) between the two groups. Age, brain natriuretic peptide (BNP) value, and PAWP (10 ± 3 vs 7 ± 3 mmHg, p < 0.001) were significantly higher in the occult LVDD group. PVR (1.7 ± 0.8 vs 2.5 ± 1.1, p = 0.020) and diastolic pulmonary pressure gradient (DPG) (− 2 ± 2 vs 1 ± 4, p = 0.004) were significantly lower in the occult LVDD group than in the non-LVDD group. Comorbidities, excluding intestinal lung disease, tended to be higher in the occult LVDD group. The proportion of participants with 3 or more co-morbidities was significantly greater in the occult LVDD group. There were no significant differences in the levels of autoantibodies, including anti-nuclear, anti-centromere, and anti-SCL70, between the 2 groups.
Haemodynamic parameters during exercise. Cardiopulmonary exercise testing with RHC findings at rest (after leg raise), AT, and peak exercise are shown in Table 2. Despite the PAWP level at rest within normal limits in both groups, its level in leg raise was significantly higher in the occult LVDD group than in the other group (15 ± 4 vs 10 ± 4 mmHg, p < 0.001). The relationships of PAWP with CO in each group are shown in Fig. 3. The mean PAWP in the occult LVDD group during exercise was higher (AT: 25 ± 5 vs 16 ± 5 mmHg, p < 0.001; peak: 28 ± 4 vs 16 ± 5 mmHg, p < 0.001), compared to PAWP in the non-LVDD group. The time constant of pulmonary circulation (RC-time) at peak exercise was significantly lower (0.23 ± 0.10 vs 0.35 ± 0.15 s, p = 0.008) in the occult LVDD group than in the non-LVDD group. CO during exercise was not significantly different between the two groups. CO at peak exercise and peak VO 2 in our study patients were 9.9 ± 4.9 L/min and 14.3 ± 8.8 mL/min/kg, respectively, indicating that their exercise capacities were low.
Echocardiographic parameters. The echocardiographic parameters between each group are shown in Table 3. There were no significant differences between the two groups in E/e' and E  (Fig. 4). A PAWP at leg raise > 11 mmHg had 100% sensitivity and 56% specificity in predicting increased PAWP ≥ 25 mmHg by exercise.

Discussion
In this cohort of SSc patients, who were mostly female (87%) and referred because of dyspnoea on exertion, onethird of the patients with normal haemodynamics at rest showed occult LVDD via exercise test. Higher PAWP after leg raise was a useful parameter for the detection of occult LVDD in these patients. PH is recognised as a critical complication which worsens prognosis in patients with SSc. This prognosis is mainly linked to cardiovascular damage 21 . Currently, there are no established therapeutic interventions for improving SSc prognosis. Although the clinical course and mortality in PAH have improved due to progresses in PAH-specific therapy, SSc-PH outcomes have not improved 22 . SSc-PH is not only caused by vasculopathy, but also causes diverse and sometimes overlapping mechanisms, including extensive vasculopathy, inflammation, autoimmunity, and widespread fibrosis 4,10 . SSc has the pathological hallmark of myocardial fibrosis and has been reported in 50-80% of necropsy cases 23 . Mechanistically, it appears that disease-related loss-offunction of small coronary arteries and arterioles leads to ischaemia and reperfusion injury, driving cardiac fibrotic tissue formation 24 . Furthermore, LV dysfunction gradually progresses to pulmonary arteriopathy and RV dysfunction 4,25 . In a previous study, at baseline, 17% of patients with SSc were diagnosed with overt LVDD evaluated by echocardiography at rest and associated with high mortality 7 . Concomitant pulmonary vascular disease and LVDD likely contribute to poor survival in patients with SSc-PH 8 . In cases of LV dysfunction, aggressive pulmonary vasodilator therapy by preload increase is somewhat counterproductive. Accurate differentiation between pre-capillary and post-capillary PH is important.
The clinical importance of pulmonary haemodynamics during exercise has been recognised in the early detection of PH 15,[26][27][28] . In SSc patients, a previous study showed that exercise PH was associated with increased mortality compared to that of normal exercise haemodynamics 29,30 . In SSc, it has been noted that borderline PH (21-24 mmHg of mean PAP) has a high likelihood of developing into overt PH 31,32 ; hence, attention has focussed on early interventions for managing borderline PH, and lowering the criteria has been advocated 33 . Exercise PH may be useful in predicting the future development of patients with borderline PH. Some patients with LVDD may have normal resting PAWP but show an abnormal increase in PAWP in response to fluid loading. PAWP 15 mmHg has been used, with high sensitivity, to distinguish between patients between pre-capillary and postcapillary PH; however, normal PAWP, especially between 12 and 15 mmHg, does not rule out the presence of LVDD 34 . Volume or exercise challenge during RHC may be useful to unmask the presence of LVDD. Fox et al. reported that several PAH patients were reclassified as having occult post-capillary PH after fluid loading, which was indicative of the difficulty of differentiation in patients with SSc 12 . It is known that PAWP in patients with SSc is more likely to be elevated than that of controls and other types of PH in fluid loading, and it is likely that occult LVDD is more common 35 .
It has been reported that 32% of SSc patients with normal resting haemodynamics and increased estimated pulmonary pressure on an exercise echocardiogram (> 40 mmHg) showed occult LVDD defined as increasing  29 . With an exercise PAWP ≥ 20 mmHg, 12 of 17 patients were reported having exercise-induced LVDD 19 . They also reported that LA dilatation might indicate occult LVDD. Our study revealed that the prevalence of occult LVDD was about one-third of SSc patients, with an exercise PAWP ≥ 25 mmHg. LA dilatation in patients with occult LVDD was also observed in our study. Haemodynamic measurements during exercise are useful for detecting early stage LVDD, but some procedural problems of examination time, invasiveness, and ergometer equipment can occur.
Fluid challenge has been reported useful in differentiating post-capillary PH 12,36 . Among HFpEF patients with normal haemodynamics at rest, a PAWP > 15 mmHg during leg raise can indicate a PAWP ≥ 25 mmHg during exercise with 91% probability 18 . These reports show that increased preload could detect left heart dysfunction. We defined occult LVDD as ≥ 25 mmHg in the supine position during exercise and showed that PAWP during leg raise was significantly higher in the occult LVDD group. In SSc patients, PAWP > 11 mmHg at leg raise may be easier for distinguishing occult LVDD than previous provocation examinations. Haemodynamic measurement with leg raise is quick, can distinguish occult LVDD easily, and may be useful for early therapeutic intervention. In our study, exercise-induced PAWP elevation was common in patients with upper limit of normal PAWP at rest. This finding is also consistent with past recommendations 34 .
Fluid challenge is an easy manoeuvre that can be performed in any catheterization laboratory without specialised equipment. Leg raise might be even easier and safer than the fluid challenge.
This study had several limitations. Our study was a retrospective, single-centre, and non-randomized trial. Our study included a limited number of cases from a single institution, and the study population did not include overt LVDD. Although we evaluated patients with normal haemodynamics at rest, BNP levels were high in both groups, and 6MWD and peak VO 2 showed low exercise capacity, which was different from that of healthy   www.nature.com/scientificreports/ subjects. Our cohort was exclusively older and comprised females; therefore, the findings of this study may not be generalisable to younger adults or males. Age is an important confounder of the pulmonary vascular response and PAWP during exercise. Differences between the occult LVDD and non-LVDD groups may be owing to differences in age 37 . There was a gap of up to 3 months between the RHC and the echocardiographic assessment, which may have affected the results. Finally, patients with pulmonary and cardiac diseases (e.g., intestinal lung disease and atrial fibrillation) were included in the cohort.
In conclusion, a high prevalence of occult LVDD was demonstrated via RHC with exercise test in SSc with normal haemodynamics at rest. The evaluation of PAWP with leg raise is easy and useful for detecting occult LVDD.     (Fig. 1). This study was approved by the Committee for Clinical Studies and Ethics of Kyorin University School of Medicine. The purposes and risks of the study were explained to the patients, who provided informed consent prior to participating. All methods were performed in accordance with the relevant guidelines and regulations.
Right heart catheterization and cardiopulmonary exercise testing. RHC was performed with a 6-F double-lumen, balloon-tipped, flow-directed Swan-Ganz catheter (Harmac Medical Products, Inc., Buffalo, NY, USA) via a transjugular approach. Baseline haemodynamic data were recorded, the zero-reference level (mid-chest) was adjusted at the start of pressure measurement, and the PAWP was obtained as the mean value of the arterial trace during occlusion. Measurements were obtained at the end of normal expirations with patients in the supine flat position to assess the right chamber, pulmonary artery pressure (mean PAP, systolic PAP, and diastolic PAP), and PAWP 34 .
An incremental symptom-limited exercise test was performed in the supine position, with an electromagnetically braked cycle ergometer (Nuclear Imaging Table with Angio Ergometer; Lode B.V.; Groningen, Netherlands) according to the ramp protocol. During cycling, the legs were elevated to approximately 30 degrees. After the parameters were stabilised, the test consisted of a 3-min rest, followed by 3 min of warmup at an ergometer setting of 10 W (60 rpm), and, finally, testing with a 1-W increase in exercise load every 6 s (totalling 10 W/min).
During the exercise, oxygen consumption (VO 2 ), carbon dioxide output (VCO 2 ), and minute ventilation (VE) were measured with a metabolic cart (Cpex-1; Inter Reha Corp., Tokyo, Japan). Prior to calculating the parameters with respiratory gas analysis, an eight-point moving average of the breath-by-breath data was obtained. Peak VO 2 was defined as the average value obtained during the last 30 s of exercise. The anaerobic threshold (AT) point was determined using the V-slope method, along with the following conventional criteria: VE/VO 2 increases after decreasing or registering as flat, whereas VE/VCO 2 remains constant or decreases 38 . The VE vs VCO 2 slope was calculated from the start of incremental exercise to the respiratory compensation point using a least-squares linear regression 39 .
Heart rate, arterial blood pressure directly recorded in the radial artery, and 12-lead electrocardiogram were monitored continuously during the test. PAP and PAWP in the RHC were measured every minute. We used the averaged mean PAP and mean PAWP during several-second periods rather than end-expiratory measurements during exercise.
Oxygen (O 2 ) saturation in arterial blood (SaO 2 ) in the radial artery and O 2 saturation in the pulmonary artery (SvO 2 ) were measured at rest, AT, and peak exercise. SaO 2 and SvO 2 were measured simultaneously. Cardiac output (CO) was determined by the Fick method using the formula: CO (L/min) = VO 2 / {1.34 × haemoglobin × (SaO 2 − SvO 2 )}. The cardiac index (L/min/m 2 ) was determined as follows: CO/body surface area (BSA). PVR, transpulmonary pressure gradient (TPG), DPG, pulmonary arterial compliance (PAC), and RC-time were calculated as: PVR (Wood units) = (mPAP − PAWP)/CO, TPG = mPAP − PAWP, DPG = diastolic PAP − PAWP, PAC = stroke volume/ pulse pressure, and RC-time = PVR × PAC. The slope of the PAWP-flow relationship (PAWP-CO slope) was calculated from the multipoint plots of the PAWP and CO using a least-squares linear regression 40 . All measurements during exercise testing were performed without supplemental oxygen.
We defined SSc with LVDD as PAWP increased during the RHC exercise test and divided the patients on this basis into two groups: occult LVDD group (PAWP ≥ 25 mmHg during exercise) and non-LVDD group (PAWP < 25 mmHg during exercise).
The 6-min walk test (6MWT) was performed according to American Thoracic Society guidelines without supplemental O 2 a day before RHC.

Echocardiography.
A transthoracic Doppler echocardiogram at the resting state was obtained and stored digitally on an Artida (Toshiba, Tokyo, Japan) or EPIQ (Philips Healthcare, Cambridge, MA USA) ultrasound system within 3 months of the RHC. Each patient was given a unique identification number to ensure that the analysis of images could be performed blinded to all invasive data and patient characteristics. The frame rate was maintained at a minimum rate of 60 frames per second. For Doppler recordings, the average of 3 to 5 consecutive beats was measured using a horizontal sweep of 75 to 100 cm/s. LV dimensions and LAD were measured from the parasternal long axis view. The LV mass was calculated and indexed to BSA. Left ventricular ejection fraction (LVEF) was calculated using Simpson's biplane method from the apical 4-and 2-chamber views. The left atrium (LA) maximal and LA minimal volumes were estimated from the apical 4-and 2-chamber views using biplane planimetry. LA maximal volume was measured from the frame just before the mitral valve opening, and LAVi was indexed for BSA 41 .
Mitral inflow was assessed in the apical 4-chamber view with the pulsed-wave Doppler sample volume placed at the tips of the mitral valve leaflets during diastole; the early (E) and late (A) peak diastolic velocities of the mitral inflow and the E wave deceleration time were thus measured. Mitral annular motion was assessed using pulsed-wave tissue Doppler with the sample volume placed in the septal (e' septal). The E/e ratio was calculated 41 .