Seismocardiography as a valuable non-exercise method for estimating peak Vo2 in cardiac patients? First experiences in Germany

Background/ Introduction: Peak oxygen consumption (pVO2) is the most important parameter for assessing cardiorespiratory fitness (CRF). Low CRF (ie., low pVO2) is associated with cardiovascular disease and all-cause mortality and therefore has considerable significance in the field of Cardiology. The gold standard method for measuring pVO2 is a cardiopulmonary exercise test (CPET). However, CPET is complex to use and is associated with increased time, costs and stress for the patients. Recently, Seismocardiography (SCG) has emerged as a new non-invasive, non-exercise method for estimating pVO2 in healthy populations. However, the validity of SCG in cardiac patients remain unclear.

Purpose: Comparison between CPET and SCG for estimating pVO2 in cardiac patients.

Methods: After a resting phase of 5 minutes in a relaxing supine position, SCG was performed. To estimate pVO2, two different algorithms were used: one for healthy populations (Seismofit v4.7.3) (trained on a data set of 436 healthy subjects) and one for cardiac patients (Seismofit cardiac) (trained on a data set of 63 cardiac subjects with ischemic heart disease [IHD] or heart failure [HF]). Following SCG, CPET was performed on a cycle ergometer on a flat ramp (10 watts/min) with continuous gas exchange measurements to determine pVO2. PVO2 was defined as the mean value of the last 30 seconds of CPET before discontinuation due to symptom-limited exhaustion.

Results: Initially, 33 patients from outpatient heart groups were investigated as described. Finally, 15 patients (male: n=14, age 73,1±6,7 yrs; BMI: 25,2±3,4; IHD: n=14; HF: n=1) could be included in the final analysis. 18 patients were excluded due to meeting exclusion criteria for the cardiac algorithm (e.g. atrial fibrillation, mechanical heart valve) or due to lack of maximum exhaustion during CPET (respiratory exchange ration [RER] <1.0). Mean CPET values during maximal effort were peak load: 136±25 watts (93% of pred.); peak heart rate: 125±26 bpm; RER at maximum exercise: 1.06±0,03; rate of perceived exertion (RPE, scale 6-20) 17.5±1.4. pVO2 during CPET was 23.5±5.7 ml/min/kg (103% of pred.). Mean values for pVO2 estimated using SCG were: 23.1±3.0 ml/min/kg for the cardiac patients algorithm (98,3% of CPET, SEE=4,3 ml/kg/min, r=0,74) (Fig.1.) and 34.1±3.8 ml/min/kg for the healthy populations algorithm (145,11% of CPET, SEE=13 ml/kg/min, r=0,31) (Fig.2.).

Conclusion: The applicability of SCG-based pVO2 estimation in cardiac patients was simple and reliable. The results of the cardiac algorithm show a clear improvement compared to the algorithm for the healthy populations, with a quite accurate agreement with CPET mean values. However, the individual deviations in individual cases are currently still (too) high for clinical purposes and related decisions. To improve accuracy and to minimize current exclusion criteria, the data base of the algorithm should be increased.