Author + information
- Piergiuseppe Agostoni, MD, PhD⁎ (, )
- Gaia Cattadori, MD,
- Anna Apostolo, MD,
- Mauro Contini, MD,
- Pietro Palermo, MD,
- Giancarlo Marenzi, MD and
- Karlman Wasserman, MD, PhD
- ↵⁎Centro Cardiologico Monzino, IRCCS, Istituto di Cardiologia, Università di Milano, via Parea 4, 20138 Milan, Italy
To the Editor:In heart failure (HF), measurement of cardiac output (CO) during exercise is important to define severity of the disease (1). Simultaneous measurements of oxygen consumption (Vo2) and CO during exercise allow calculation of arterio-venous oxygen difference [C(a-v)o2] (2), and, plotting these three variables together, to discriminate exercise limitations due to altered left ventricle pump function from those due to other causes, including muscle enzyme deficiency and deconditioning.
The ideal method for determining CO during exercise should be non-invasive. Inert gas rebreathing (R) with continuous analysis of respired gases is a reliable, safe, and inexpensive method for noninvasive measurements of pulmonary blood flow (PBF), which is equivalent to CO in the absence of shunts.
This study was undertaken in HF patients to assess reliability and repeatability of CO measured during exercise by R using a new device with photoacoustic analyzer instead of mass-spectrometer. We compared CO measurements by R with CO by direct Fick (F) method and thermodilution (T) method. Using the data obtained noninvasively, we constructed the CO/C(a-v)o2/Vo2plot to determine the pathway leading to exercise intolerance in HF patients.
Twenty chronic HF patients (18 males and 2 females, age 53 ± 12 years, New York Heart Association functional class I in 3 cases, class II in 16 cases and class III in 1 case, sinus rhythm 17 cases, atrial fibrillation 3 cases) in stable clinical condition participated in the study. All subjects provided written informed consent to the study.
The direct Fick method (F): CO = Vo2/C(a-v)o2. We used the mean of the Vo2recorded in the last 2 min of each step. The C(a-v)o2was calculated from blood samples collected simultaneously from pulmonary and systemic arteries and immediately measured.
For the thermodilution method (T), we injected five times at each exercise step 10 ml of iced saline via a 7-F thermodilution Swan Ganz catheter into the right atrium. Reported CO data are the mean of the three closest measurements.
For the inert gas rebreathing method (R), we used N2o(blood soluble gas) and SF6 (blood insoluble gas), with concentrations, enriched with o2, of 0.5% and 0.1%, respectively (3) (Innocor, Innovision A/S, Odense, Denmark). Tidal volume was progressively increased in the closed circuit to match the physiologic increase. Use of SF6allowed us to measure the volume of lungs, valve and rebreathing bag. N2oconcentration decreases during the rebreathing maneuver, with a rate proportional to PBF. Three to four respiratory cycles were needed to obtain N2owashout. Absence of pulmonary shunt was defined as arterial o2saturation >98% (blood samples obtained from the arterial line). In the absence of pulmonary shunt, PBF = CO. In the presence of shunt PBF = CO + shunt flow (see Appendix).
The first CPET was performed on a cycle ergometer, work rate was increased in a ramp pattern selected to achieve peak exercise in 10 min. As for the second and third CPET, gas exchange and R systems were in series. Four-minute step increments equal to one-quarter of the peak exercise workload were used. In the second CPET CO was measured by R. In the third CPET, CO was measured by R, T, and F. We measured CO by T, R, and F, always in that order, at rest, and after the second minute of each workload step.
Patients had moderate HF (peak Vo2= 16.6 ± 2.9 ml/min/kg, ejection fraction 40 ± 9% by echocardiography). The mean workload increment for each step in the second and third CPET, was 25 ± 7 W. Table 1reports hemodynamic parameters, heart rate, Vo2, ventilation, tidal volume, hemoglobin, arterial hemoglobin o2saturation, CO measurements obtained with T, F, and R, and shunt flow at each step of the third CPET. A significant functional shunt (oxyhemoglobin saturation <98%) was frequently detected. Repeatability of PBF measurements by R was assessed by comparing the results of CPET 2 and 3 (Table 1). The coefficient of variation was 10.8%.
We performed 95 simultaneous CO measurements by the 3 techniques. CO by T versus F was: COT = 1.1425 × COF – 1.358; R = 0.9465; R versus F was: COR = 0.9504 × COF + 0.5366; R = 0.9482; R versus T was: COR = 0.779 × COT + 2.1037; R = 0.9382. C(a-v)o2can be estimated from measured Vo2and CO by R. The estimated C(a-v)o2was highly correlated to the measured value: C(a-v)o2estimated = 0.9326 × C(a-v)o2measured + 0.5598; R = 0.9040. The Bland-Altman plot confirmed the good agreement among the three techniques. No statistical differences were observed between CO measurements by any technique nor between estimated versus measured C(a-v)o2.
In Figure 1CO is plotted versus C(a-v)o2. The solid lines are isoVo2lines. Full symbols are data from F (measured Vo2, measured C(a-v)o2and calculated CO); open symbols are data from R (measured CO, measured Vo2, and estimated C(a-v)o2).
It is important to measure CO during exercise in HF. Indeed, a relevant number of patients have a good prognosis, despite low peak Vo2, if CO increase during exercise is preserved. At present, no data regarding the use of CO response to exercise as a tool to assess patients’ follow-up are available. This is due to the invasive nature of CO determination during exercise.
The R, T and F CO measurements provided similar results, signifying that both PBF measurements and shunt estimation are reliable. These results were obtained in patients with moderate HF. Stringer et al. (2) reported the CO/C(a-v)o2/Vo2plot of both healthy and HF subjects. For a given Vo2, HF patients have a lower CO and a greater C(a-v)o2, when compared to normal subjects. This plot should differentiate among those patients who have muscle deconditioning from those who are more fit. This is important for the selection of the most appropriate patients for intensive cardiac rehabilitation programs. Indeed, the reduced CO response to exercise may not be the sole limitation to physical activity in heart failure patients. The present study shows that this plot can be noninvasively built because the estimation of C(a-v)o2by R-CO and Vo2measurements appear to give reliable values for C(a-v)o2.
In conclusion, CO in HF can be measured during exercise by R. These measurements are repeatable and agree closely with those from F and T. We showed that C(a-v)o2can be estimated and the CO/C(a-v)o2/Vo2plot can be built using measurements obtained noninvasively.
Shunt Flow Calculation
(a)(b)(c)Rearranging equation bgives:Inserting into equation cgives:or(d)Insertion of equation din equation agives us the shunt flow equation:Vo2, CO, PBF, Cao2and Cvo2were directly measured, while Cco2was estimated by assuming pulmonary capillary o2saturation = 98% and using hemoglobin values obtained from the arterial blood samples.
Please note: This paper was sponsored by a Centro Cardiologico Monzino Research Grant.
- American College of Cardiology Foundation
- Metra M.,
- Faggiano P.,
- D’Aloia A.,
- et al.
- Stringer W.W.,
- Hansen J.E.,
- Wasserman K.
- Kallay M.C.,
- Hyde R.W.,
- Smith R.J.,
- Rothbard R.L.,
- Schreiner B.F.