Feasibility of the cardiac output response to stress test in suspected heart failure patients

Abstract Background Diagnostic tools available to support general practitioners diagnose heart failure (HF) are limited. Objectives (i) Determine the feasibility of the novel cardiac output response to stress (CORS) test in suspected HF patients, and (ii) Identify differences in the CORS results between (a) confirmed HF patients from non-HF patients, and (b) HF reduced (HFrEF) vs HF preserved (HFpEF) ejection fraction. Methods Single centre, prospective, observational, feasibility study. Consecutive patients with suspected HF (N = 105; mean age: 72 ± 10 years) were recruited from specialized HF diagnostic clinics in secondary care. The consultant cardiologist confirmed or refuted a HF diagnosis. The patient completed the CORS but the researcher administering the test was blinded from the diagnosis. The CORS assessed cardiac function (stroke volume index, SVI) noninvasively using the bioreactance technology at rest-supine, challenge-standing, and stress-step exercise phases. Results A total of 38 patients were newly diagnosed with HF (HFrEF, n = 21) with 79% being able to complete all phases of the CORS (91% of non-HF patients). A 17% lower SVI was found in HF compared with non-HF patients at rest-supine (43 ± 15 vs 51 ± 16 mL/beat/m2, P = 0.02) and stress-step exercise phase (49 ± 16 vs 58 ± 17 mL/beat/m2, P = 0.02). HFrEF patients demonstrated a lower SVI at rest (39 ± 15 vs 48 ± 13 mL/beat/m2, P = 0.02) and challenge-standing phase (34 ± 9 vs 42 ± 12 mL/beat/m2, P = 0.03) than HFpEF patients. Conclusion The CORS is feasible and patients with HF responded differently to non-HF, and HFrEF from HFpEF. These findings provide further evidence for the potential use of the CORS to improve HF diagnostic and referral accuracy in primary care.

myocardial infarction and worsening HF. 7 The prevalence of HF and consequential hospitalization are expected to rise due to our ageing population 8 affecting ≥10% in individuals >70 years of age. 1 Typical signs and symptoms of HF include breathlessness, ankle swelling, and fatigue, 1 which are very common and nonspecific. This makes diagnosis challenging due to difficulties in differentiating between HF and other conditions associated with ageing, obesity, or lung disease. 1,9 In addition, multimorbidity and polypharmacy can complicate the diagnosis further. 10 Early and precise diagnosis of HF is therefore essential in order to apply appropriate treatment, which can improve patient's morbidity, quality of life, and mortality 11 as well as reducing NHS burden of HF. 1,2,12 Primary care represents an entry point in the clinical care pathway for patients presenting with HF signs and symptoms and general practitioners (GPs) play a key role in identifying HF. 11,13,14 Currently, the diagnostic tools available in primary care to help diagnose HF include a full medical history, signs and symptoms, electrocardiography, and serum natriuretic peptides test. 1 However, electrocardiography seems to be an inadequate screening tool for patients with suspected HF as described by a systematic meta-analysis. 15 The serum natriuretic peptides blood test is useful as a "rule out" test but results in a high number of false positives. 11,[16][17][18] Furthermore, this diagnostic test continues to be underutilized by GPs in primary care. 19 In practice, all these assessments are used to establish a HF diagnosis, which also therefore requires some specialist clinical interpretation. It is therefore not surprising that current diagnostic practice of HF in primary care results in high number of inaccurate and expensive referrals to secondary care. 15,[20][21][22] Diagnostic uncertainty and delayed diagnosis of HF can lead to incorrect treatment, impaired disease progression and increased hospitalization. 11 The COVID-19 pandemic has further reinforced that new strategies to achieve timely diagnosis due to patients avoiding urgent care for HF signs and symptoms 23,24 in primary care should be a priority for future research and policy as most recently suggested. 25 To help mitigate this challenge, we have developed and recently confirmed acceptable reproducibility of a novel, easy-to-administer, noninvasive cardiac output response to stress (CORS) test 26 (Fig. 1). Our recent qualitative work with healthcare professionals in primary and secondary care highlighted a high demand for an additional diagnostic tool within primary care to aid diagnostic and referral accuracy. 27 We hypothesis that the CORS test may enable a new clinical pathway-once the serum natriuretic peptides blood test has been reported, GPs would refer patients for the CORS test who demonstrated elevated levels of natriuretic peptides. 1 This has the potential to overcome the reduced specificity of the blood test 11,17,18 and improve referral accuracy. The purpose of this present study was: (i) to determine if patients suspected of HF were able to complete the CORS test after being referred to secondary care and (ii) to evaluate differences in the CORS test parameters between HF and non-HF, and heart failure reduced (HFrEF) vs heart failure preserved (HFpEF) ejection fraction.

Study design, setting, and patients
In this single centre, prospective, observational, feasibility study, 105 consecutive suspected HF patients were recruited from HF diagnostic clinics at the Royal Victoria Infirmary, Newcastle upon Tyne Hospitals National Health Service Foundation Trust, Newcastle upon Tyne, United Kingdom. Data were collected between February 2017 and March 2020. Patients aged ≥50 years were included in the study if they were suspected of having HF by their GPs (based on symptoms, signs, physical examination, and raised natriuretic peptides levels) and had been referred to secondary care to a HF diagnostic clinic for specialist review. Signs and symptoms included

Key messages
• Suspected heart failure patients were able to complete the CORS test.
• Different results were found between heart failure and nonheart failure patients. • Differences were found between ejection fraction groups.  breathlessness, ankle swelling, and fatigue, which are found in the ESC HF guidelines. 1 Exclusion criteria included a previous diagnosis of HF, the presence of severe symptoms requiring urgent assessment and stabilization, recent acute coronary syndrome and severe physical disability. The study protocol (number 16/NE/0287) was approved by the National Health Service, National Research Authority (North East-Tyne & Wear South Research Ethics Committee). All procedures performed in the study were in accordance with the Declaration of Helsinki. Patients gave written informed consent.

Study protocol and measurements
The CORS test has been described previously. 26 It uses continuous haemodynamic measurements based on noninvasive, electrical signal processing bioreactance technology (NICOM, Cheetah Medical, Inc., MA, USA), which we have previously validated. [28][29][30] The CORS test has demonstrated acceptable reproducibility in healthy adults (≥50 years old) 26 and consists of 4 phases: (i) rest (patients lies in the supine position for 3 min); (ii) challenge (patients remains in the standing position for 3 min); (iii) and (iv) stress-step exercise (patient completes a step test on a 15-inch step board with intensity controlled by a metronome at 10 and 15 steps per min). 26 Due to time restrictions in the clinic, patients in this study only completed 1 stress-step exercise phase at the 15 steps per min intensity.
Suspected HF patient NTproBNP (ng/L) levels were recorded from patient records, which the consultant cardiologist would extract as part of the diagnostic tests to confirm or refute a diagnosis of HF. NICE guidelines recommend referral to specialist assessment and echocardiography when NTproBNP levels are between 400 and 2,000 ng/L within 6 weeks whereas levels >2,000 ng/L require referral within 2 weeks. 2 Patients completed a number of diagnostic tests including echocardiography (estimated left ventricular ejection fraction, LVEF %) and electrocardiography, and were seen by a consultant cardiologist-HF specialist who reviewed their medical history, diagnostic test results, and conducted a physical examination to confirm or rule out a diagnosis of HF. The consultant cardiologist confirmed a diagnosis of HFrEF by estimated LVEF <40% or HFpEF by estimated LVEF >50%. There were 6 patients diagnosed with HF with midrange ejection (HFmrEF) who were grouped with HFrEF for the purposes of this study. All CORS test assessments for this study were administered by a researcher and completed after the patient had been seen by the consultant cardiologist and a diagnosis of HF was confirmed or refuted. The researcher was blinded from echocardiology and patient diagnosis at the time of conducting the CORS test.

Statistical analysis
Data were analysed using SPSS, version 25 (SPSS, Inc., Chicago, IL, USA). The level of significance was set at P <0.05. Data are expressed as mean ± SD unless otherwise indicated. Prior to statistical analysis, data were screened for univariate outliers using standard Z-distribution cut-offs. Shapiro-Wilk test was used to assess normality of distribution and Levene's test confirmed homogeneity of variances. Differences between the 2 groups (confirmed HF and non-HF patients) were assessed using either an independent t-test or Mann-Whitney U test. In a separate subset analysis, confirmed HF patients only were grouped by ejection fraction diagnosis (HFrEF vs HFpEF) and differences were assessed using either an independent t-test or Mann-Whitney U test.

Results
Patients' demographics, clinical characteristics, and medication list are presented in Table 1. Confirmed HF patients had a significantly lower estimated LVEF % (43 ± 12% vs 54 ± 4%, P < 0.01) and higher NTproBNP (1,883 ± 1,957 vs 748 ± 910 ng/L, P < 0.01) than non-HF patients. There were no other significant differences between the groups in their demographic and clinical characteristics. All patients completed the rest phase of the CORS test (38 with HF and 67 without HF). While all confirmed HF patients were able to complete the challenge phase, this was the case for 93% of non-HF cases. A total of 79% of newly diagnosed HF patients completed all 3 phases of the CORS test, including the stress-step exercise, whereas 91% of non-HF patients completed all 3 phases of the CORS test. The reasons for missing measurements varied. The consultant cardiologist requested that the patient not complete either the challenge-standing (non-HF patients, n = 1) and/or the stress-step exercise (confirmed HF patients, n = 4; non-HF patients, n = 2) phases due to their condition or there was poor connectivity and signal loss during measurements (challenge-standing phase: non-HF patients, n = 4; stress-step exercise phase: confirmed HF patients, n = 2; non-HF patients, n = 3) or the patient reported balance issues for the stress-step exercise phase and researcher did not complete this phase (confirmed HF patients, n = 2; non-HF patients, n = 1). For the challengestanding and stress-step exercise phases that included complete patient haemodynamic measurements but resulted in missing blood pressure measurements were due to poor connectivity or system time out from the bioreactance technology as the patient was not seated at the time of these phase measurements (missing blood pressure measurements: 5% for challenge-standing phase for confirmed HF patients, n = 2/38; 23% for stress-step exercise phase for both confirmed HF patients, n = 7/30 and non-HF patients, n = 14/61). Table 2 displays the haemodynamic variables for patients with confirmed HF and non-HF for the 3 phases of the CORS test (rest, challenge, and stress-step exercise). At rest, patients with confirmed HF had a significantly higher heart rate (74 ± 16 vs 68 ± 16 beats/min, P = 0.03), lower stroke volume (84 ± 28 vs 97 ± 33 mL/beat, P = 0.04), lower stroke volume index (SVI; 43 ± 15 vs 51 ± 16 mL/beat/m 2 , P = 0.02), and higher diastolic blood pressure (85 ± 12 vs 80 ± 11 mm Hg, P = 0.02) than non-HF patients.
For the challenge phase of the CORS test, heart rate was significantly higher in the confirmed HF patients (81 ± 18 vs 72 ± 14 beats/min, P = 0.01) but no other variables were significantly different between confirmed HF and non-HF. Patients with confirmed HF had a significantly lower SVI compared with the non-HF group for the stress-step exercise phase of the CORS test (49 ± 16 vs 58 ± 17, mL/beat/m 2 , P = 0.02).

Discussion
The purpose of the present study was to determine the feasibility for completing the CORS test in suspected HF patients and to evaluate the differences in the CORS test parameters between HF and non-HF and patients with HFrEF and HFpEF. The preliminary findings highlight that nearly 80% of newly diagnosed HF patients were able to successfully complete all phases of the CORS test (including step exercise). There were differences in haemodynamic function between confirmed HF and non-HF patients at each of the phases of the CORS test. There were further differences between patients with HFrEF and those with a diagnosis of HFpEF.
The ESC HF guidelines defines HF as a clinical syndrome associated with cardiac functional and/or structural abnormalities resulting in reduced cardiac output at rest and/or during stress. 1 However, in the description of diagnostic pathway there is no indication about monitoring and reporting haemodynamic measurements in patients undergoing HF diagnostic  25 The present study reveals that the majority of patients were able to complete all phases of the CORS test (rest-supine, challenge-standing, and stress-step exercise). This was a concern raised by healthcare professionals during our qualitative evaluation of the potential implementation of the CORS test for suspected HF in primary care. 27 The CORS test was administered in 18 ± 3 min so an additional appointment will be required for this test to be included within primary care. The qualitative work highlighted that not all phases of the test would be suitable for all patients. 27 This current study has seen adaptations to the CORS test such as reducing the 2-phase stress-step exercise (originally proposed 26 ) to 1 phase to be time efficient and using a balance aid for the stress-step exercise phase when the patient required, which are important developments if as suggested from our qualitative work that either a practice nurse or healthcare assistant would be able to deliver the test within a primary care setting. 27 Using such approach we achieved a 93% and 79% success rate for completion of the stress-exercise CORS test in patients with suspected HF and those with confirmed HF. This finding addresses previous concerns identified in our qualitative work by GPs and practice nurses that completion of the CORS test will be achieved in only 50%-80% of patients. 27 It should also be noted that haemodynamic measurements during the rest phase of the CORS test could be achieved in all patients, and can yield important information about the presence and type of HF, as described below.
In patients with confirmed HF, heart rate was significantly higher at both rest and challenge phases compared with non-HF patients. Likewise, for all 3 phases of the CORS test, heart rate was higher in patients with HFrEF compared with patients with HFpEF. In patients with HFrEF, findings highlight that, for every beat increase in heart rate, there is a 3% increase in risk of mortality or hospital admission for worsening HF. 31 Heart rate control is a vital part of HF management. 31,32 As the patients in our study were recruited from a diagnostic clinic, this was before treatment optimization.
Both the rest and stress-step exercise phases resulted in a lower stroke volume and/or SVI for confirmed HF patients Lower stroke volume has been reported previously in patients with HFrEF at rest and at exercise compared with patients with HFpEF. 33 Similarly, lower SVI has been found at rest in patients with HFrEF (n = 157) compared with healthy participants (n = 147) using cardiovascular magnetic resonance. 34 Diastolic blood pressure was significantly different at rest with a higher reading recorded in the confirmed HF group (vs non-HF group). No other significant blood pressure differences were recorded between groups. Although, in the subset analysis, systolic blood pressure was lower in the HFrEF group at the challenge phase compared with the HFpEF group. Interestingly, in the confirmed HF group, systolic and diastolic blood pressure decreased from supine to challenge phases. Abnormal haemodynamic postural response differences have been found previously in HF patients with lower blood pressure recorded at a standing position vs supine, which may present as postural hypotension. 35 For the first time, we have provided results from suspected HF patients completing the CORS test. However, this study is not without limitations. This was a single centre, prospective observational, feasibility study with moderate sample size.
Our ongoing research programme aims to confirm the diagnostic accuracy (sensitivity and specificity) of the CORS test with the objective of its implementation in primary care to improve diagnosis of HF. Although at the time of performing the CORS test, all patients were suspected of having HF, Table  1 shows differences in medication used at the time of investigations, which may influence haemodynamic function. We acknowledge that at the point of HF diagnosis that medication was not titrated. We did not use a visual analogue scale to assess patient burden of completing the CORS test, which would have strengthened our feasibility findings. The patient was not blinded from their diagnosis at the time of completing the CORS test and we acknowledge that this may have impacted on their well-being and capacity to perform the test. The patients were seen by one of 3 consultant cardiologists (HF specialists), however, we did not complete inter-rater variability between the clinicians regarding HF diagnosis.
The major findings of the present study highlight that the CORS test can be completed by the majority of patients undergoing a clinical investigation for a diagnosis of HF within a secondary care setting. The findings have shown that performing the CORS test in HF suspected patients is feasible within a secondary care setting. Furthermore study results demonstrate haemodynamic differences between HF and non-HF as well as HFrEF vs HFpEF. It appears that SVI at rest and its response to challenge (standing) and/or short step exercise differentiate the most between patients with HF from non-HF patients and HFrEF from HFpEF. Our ongoing research will define cut-off values for the CORS test variables and confirm their sensitivity and specificity to complement diagnosis of HF in primary care.