Chemoreflex sensitivity as a predictor of arrhythmia relapse in ICD recipients

International Journal of Cardiology 86 (2002) 169–175 www.elsevier.com / locate / ijcard

Chemoreflex sensitivity as a predictor of arrhythmia relapse in ICD recipients Marcus G. Hennersdorf*, Verena Niebch, Christian Perings, Bodo E. Strauer Department of Cardiology, Pneumology and Angiology, Medical Clinic and Policlinic B, Heinrich-Heine-University, Moorenstr. 5, D-40225 Dusseldorf, Germany Received in revised form 8 April 2002; accepted 5 May 2002

Abstract Background: The chemoreflex sensitivity as a marker of a disturbed vagal reflex activity has proved to be a parameter of increased risk for ventricular tachyarrhythmias or sudden cardiac death. The sensitivity of patients with prior myocardial infarction concerning ventricular tachyarrhythmias amounted to about 70%. This prospective study should evaluate the positive predictive accuracy of this new method in patients at risk for ventricular arrhythmias. Methods: 42 patients were enrolled into this study. All had a prior myocardial infarction at least 6 months previously; 35 patients were resuscitated from sudden cardiac death, and seven patients had documented monomorphic ventricular tachycardias. All patients were recipients of an ICD. The chemoreflex sensitivity was measured by determination of the venous partial pressure of oxygen and the heart rate before and after inhalation of pure oxygen. The difference in the RR-intervals before and after inhalation divided by the difference in the oxygen pressures were calculated as the chemoreflex sensitivity [ms / mmHg]. Furthermore, in all patients additional risk stratifiers used in this study were the presence of ventricular late potentials (LP), the short-term heart rate variability (HRV), the baroreflex sensitivity (BRS) and a decreased left ventricular function (ejection fraction,40%, EF). Results: The chemoreflex sensitivity in the patient group as a whole amounted to 2.5962.06 ms / mmHg. During follow-up, out of the 42 patients enrolled, 20 had a documented arrhythmic event (AE: sustained ventricular tachycardia or ventricular fibrillation). Patients with and without AE showed significantly different values of chemoreflex sensitivity (1.5861.09 vs. 3.5162.31 ms / mmHg, P,0.01) and EF (33.3615.6 vs. 47.9617.9%, P,0.05), but not of LP, HRV or BRS. The relative risk of reduced chemoreflex sensitivity concerning an AE amounted to 2.83 (95% CI 0.99–8.01). Conclusions: The chemoreflex sensitivity as a marker of increased risk for ventricular tachyarrhythmias shows a high positive predictive power in patients with prior myocardial infarction and who previously survived ventricular tachyarrhythmias. These results should be confirmed by studies in broad populations and without survived arrhythmic event.  2002 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Chemoreflex sensitivity; Ventricular arrhythmia; Myocardial infarction

1. Introduction For the prediction of arrhythmic events in patients with coronary artery disease and a prior myocardial *Corresponding author. Tel.: 149-211-811-8800; fax: 149-211-819520. E-mail address: [email protected] (M.G. Hennersdorf).

infarction, apart from the analysis of ventricular late potentials [1,2] and repolarization abnormalities [3,4] the analysis of a disturbed sympathovagal balance is helpful. Heart rate variability [5,6] and baroreflex sensitivity [7,8] are known methods to determine a decreased vagal tone and vagal reflex activity, respectively. A correlation between a decreased vagal tone and ventricular arrhythmias exists, but the positive

0167-5273 / 02 / $ – see front matter  2002 Elsevier Science Ireland Ltd. All rights reserved. PII: S0167-5273( 02 )00191-2

170

M.G. Hennersdorf et al. / International Journal of Cardiology 86 (2002) 169–175

predictive accuracy of these methods for the prediction of life threatening arrhythmias amounts to a maximum of only 50% [9]. Hence, further methods are necessary to stratify patients at risk with a higher predictive accuracy. It is well known that oxygen increase leads to a decrease in heart rate [10]. This effect results from a deactivation of peripheral arterial chemoreceptors [11] that are located in the glomus caroticum (oxygen) and area postrema of the middle brain (carbon dioxide). After integration of the neural signal in the medulla oblongata a decrease in the heart rate occurs caused by a vagal activation. But the degree of heart rate reduction due to oxygen inhalation is different in patients with and without survived sudden cardiac death [12]. Patients who survived sudden cardiac death are characterized by a significantly smaller decrease in the heart rate after oxygen inhalation than patients without ventricular tachyarrhythmias. The sensitivity of a decreased chemoreflex sensitivity concerning a survived severe arrhythmic event amounted to 71% in a prior study [12]. The aim of the present study was to evaluate the predictive accuracy of this new method in a population with increased risk for life threatening ventricular arrhythmias.

2. Methods

2.1. Patient population In this study, 42 consecutive patients with coronary artery disease and previous myocardial infarction were enrolled. Out of these, 35 patients were survivors of sudden cardiac death and seven patients suffered from sustained ventricular tachycardia (Table 1). All patients were supplied with an implantable cardioverter / defibrillator (CPI Ventak P, Ventak PRx and Ventak Mini, Guidant, USA). Exclusion criteria were atrial flutter or fibrillation and obstructive pulmonary disease. All patients were followed-up for up to 24 months. The end-point of this study was defined as an arrhythmic event like ventricular fibrillation, ventricular flutter or sustained ventricular tachycardia, that required therapy by antitachycardiac pacing or

Table 1 Patient characteristics Parameter

Patients

Gender (M / F) Age (years) Ejection fraction (%) 1-vessel CAD (%) 2-vessel CAD (%) 3-vessel CAD (%) Previous MI Previous SCD (%) Previous sustained VT (%) NYHA class (I / II / III)

28 / 14 61.3610.5 41.5618.2 13 (31) 11 (26) 18 (43) 42 (100) 35 (83) 7 (17) 8 / 19 / 15 (19 / 45 / 36)

CAD, coronary artery disease; MI, myocardial infarction; SCD, sudden cardiac death; VT, ventricular tachycardia.

shocking of the implanted ICD. The patients were seen as outpatients every 3 months.

2.2. Measurement of chemoreflex sensitivity The patients were resting over a period of 10 min. After this phase, a blood sample from a cubital vein was taken and the partial oxygen pressure was determined using a standardized blood gas analyzer (Radiometer, Copenhagen). Additionally, the mean RR-interval out of 10 consecutive RR-intervals was calculated using a standardized 12-channel electrocardiogram (Siemens). Then, the patients inhaled 100% pure oxygen (5 l / min) via a nasal mask. No conversation was carried out during this period for minimization of mental influences. After a period of 5 min, the venous partial oxygen pressure and the mean RR-interval out of 10 consecutive RR-intervals were determined again. The difference between the mean RR-interval before and after oxygen inhalation divided by the difference between venous partial oxygen pressure before and after oxygen inhalation was treated as the chemoreflex sensitivity (ChRS) (ms / mmHg). Every patient gave written informed consent; there were no objections from the ethical point of view. According to previous papers, a chemoreflex sensitivity below 3.0 ms / mmHg (representing the median value) was considered as pathologically decreased [12].

2.3. Invasive investigation All patients were investigated by invasive left heart

M.G. Hennersdorf et al. / International Journal of Cardiology 86 (2002) 169–175

catheterization and coronary angiography. In case of suspected elevated right heart pressures a right heart catheterization was performed additionally. Determination of ejection fraction was performed by analysis of biplane ventriculographies using the Dodge formula [13]. In all 42 patients, an electrophysiologic study was carried out. Up to three extrastimuli were delivered from up to two right ventricular sites. Electrophysiologic tests were considered positive if sustained ventricular tachycardia or ventricular fibrillation were inducible. Other investigated non invasive risk stratifiers were baroreflex sensitivity, heart rate variability and late potential analysis. Baroreflex sensitivity was measured by analyzing the decrease in the heart rate following the injection of a selective agonist of the alpha-1-adrenoceptor (phenylephrine) as previously described by other authors [14,15]. A test was considered as pathologically decreased, if the value was ,3.0 ms / mmHg [7,8]. The measurement of heart rate variability was carried out over a period of 15 min (Predictor, Corazonix, USA). It was defined as pathologically decreased, if the S.D. of all consecutive RR-intervals was ,30 ms, as previously reported [16]. The analysis of ventricular late potentials was performed using a signal-averaged ECG (Predictor, Corazonix, USA) recording orthogonal X,Y,Z leads. Approximately 200 ECG cycles were digitally recorded until a noise level of ,0.5 mV was obtained. Late potential criteria were the duration of the filtered QRS-complex (.114 ms), the root mean square amplitude of the last 40 ms (,20 mV) and the low amplitude signals at the end of the QRS-complex below 40 mV (.38 ms). If two or three parameters were met, late potentials were classified as being present [2].

171

tic-regression model to identify independent predictors of ventricular arrhythmias during follow-up. The role of chemoreflex sensitivity in predicting arrhythmia-free survival was evaluated with the log-rank test. Data are expressed as mean6S.D.

3. Results All patients were followed-up for up to 24 months. The median follow-up time amounted to 12 months with a range of 1–24 months.

3.1. Characteristics of chemoreflex sensitivity All patients showed a significant increase in the partial oxygen pressure (40.069.9 vs. 52.2616.8 mmHg, P,0.0001) and the RR-intervals (843.16185.6 vs. 861.96185.3 ms, P,0.0001) during oxygen inhalation. The chemoreflex sensitivity of the patient group as a whole amounted to 2.5962.06 ms / mmHg. The chemoreflex sensitivity was significantly higher in patients with NYHA class I compared to those with NYHA class II or III (Fig. 1). The baseline values of patients with and without decreased chemoreflex sensitivity were not significantly different (RR-interval: 880.66195.7 vs. 777.26159.1 ms, n.s.; pO 2 : 39.8610.6 vs. 40.069.3 mmHg, n.s.).

2.4. Statistical analysis Parametric and nonparametric tests were used when appropriate. Yates correction was used for the x 2 -test. In addition to the chemoreflex sensitivity, relevant clinical variables (age, gender, NYHA-classification, left ventricular ejection fraction ,40%) and other noninvasive risk stratifiers (ventricular late potentials, heart rate variability, baroreflex sensitivity) were analyzed with a multivariate stepwise logis-

Fig. 1. Significantly different values for the chemoreflex sensitivity of patients with NYHA class I compared to those with NYHA class II or III.

172

M.G. Hennersdorf et al. / International Journal of Cardiology 86 (2002) 169–175

The chemoreflex sensitivity and baroreflex sensitivity were correlated significantly on a poor level (r50.31, P50.05). The values of CO 2 partial pressure did not change significantly (43.164.5 vs. 43.863.9, P5n.s.).

3.2. Characteristics of inducible arrhythmias In all 42 patients, an electrophysiologic study was performed; 62% of these patients had pathological electrophysiological results (12 sustained ventricular tachycardia and 14 ventricular fibrillation). The mean cycle length of the induced ventricular tachycardias was 322.3699.7 ms.

3.3. Correlation of electrophysiologic study and chemoreflex sensitivity Patients with a reduced chemoreflex sensitivity were inducible in the electrophysiologic study in 72%. Patients with and without inducible ventricular arrhythmias in the electrophysiologic study showed a different chemoreflex sensitivity, but without reaching the level of significance (2.3461.89 vs. 3.2862.42 ms / mmHg in patients without inducibility, P5n.s.).

3.4. Follow-up During the follow-up period of at most 24 months, 20 patients (48%) had an arrhythmic event. Out of these patients 16 (80%) showed symptomatic ventricular tachycardia and four patients showed ventricular fibrillation or flutter. No patient died during the follow-up period. Among patients with an arrhythmic event during follow-up, the chemoreflex sensitivity was significantly different between patients with and without an event (1.5861.09 vs. 3.5162.31 ms / mmHg, P, 0.001). The positive predictive value of a decreased chemoreflex sensitivity amounted to 61%, the sensitivity to 85%, the negative predictive value to 85%, the specificity to 50% and the relative risk to 2.83 (95% CI 0.99–8.01) (P,0.01). The left ventricular ejection fraction was significantly lower in patients with an arrhythmic event, and those patients were more often inducible in the electrophysiologic study (Table 2). No significant differences concerning baroreflex sensitivity, late potential analysis and heart rate variability could be evaluated (Table 3). Although the patients with arrhythmic event were more often treated by ace-inhibitors and betablockers, the chronic medication showed no significant differences between both groups. The logistic regression analysis showed that in the

Table 2 Characteristics of patients with and without an arrhythmic event during follow-up and univariate analysis of risk stratifiers Parameter

Event (n520)

No event (n522)

P-value

Age (years) Gender (m / f) NYHA (I / II / III / IV) Ejection fraction (%) Ejection fraction ,40% (%) EPS inducible (%) ChRS,3.0 ms / mmHg (%) BRS,3.0 ms / mmHg (%) Late potential positive (%) HRV,30 ms (%) Betablockers (%) ACE-inhibitors (%) Diuretics (%) Antiarrhythmic drugs (%) Sotalol Amiodarone Class I

64.469.8 15 / 5 2 / 10 / 8 / 0 33.3615.6 60 95 85 53 56 82 60 85 45

58.8610.5 13 / 9 6/9/7/0 47.9617.9 32 8 50 23 36 67 45 64 36

n.s. n.s. n.s. ,0.05 0.06 ,0.01 ,0.05 ,0.05 n.s. n.s. n.s. n.s. n.s.

10 20 0

ChRS, chemoreflex sensitivity; BRS, baroreflex sensitivity.

9 9 0

n.s. n.s. n.s.

M.G. Hennersdorf et al. / International Journal of Cardiology 86 (2002) 169–175

173

Table 3 Statistical values for the tested variables Parameter

Sensitivity

Specificity

PPV

NPV

RR

ChRS (,3 ms / mmHg) BRS (,3 ms / mmHg) HRV (,30 ms) LP (1) EF,40%

85 53 82 56 60

50 77 33 64 68

61 67 50 56 63

79 65 70 64 65

2.82 1.93 1.66 1.55 1.82

PPV, NPV, positive, negative predictive value; RR, relative risk.

model including clinical variables, the only independent predictor of arrhythmia relapse was the chemoreflex sensitivity. In the model including other non invasive risk factors, such as baroreflex sensitivity and left ventricular function, the chemoreflex sensitivity and a reduced left ventricular ejection fraction (,40%) were independent predictors of a ventricular tachyarrhythmia (Table 4). Kaplan–Meier analysis (Fig. 2) showed significantly different arrhythmia free survival rates for patients with and without decreased chemoreflex sensitivity, but not for patients with or without decreased baroreflex sensitivity. The cumulative arrhythmia free survival rate of patients with a left ventricular ejection fraction .40% was higher than of those ,40%, but did not reach the level of significance (P50.11).

4. Discussion In the last decade, many noninvasive methods for the identification of patients at risk for ventricular arrhythmias were reported. Regarding a decreased vagal tone, in 1987, Kleiger et al. [5] demonstrated that a decreased heart rate variability is associated with an increased risk for arrhythmic events. Newer studies confirm these findings [17]. Already in the 1970s [18], but in particular in the last 10 years, the

Fig. 2. Arrhythmia free survival in 42 patients according to the results of chemoreflex sensitivity (top) and baroreflex sensitivity testing (middle) and the left ventricular function (bottom, LV.40, LV,405left ventricular ejection fraction . or ,40%).

analysis of the baroreflex sensitivity was evaluated showing a high predictive accuracy regarding ventricular tachyarrhythmias for patients with survived

Table 4 Multivariate logistic regression analysis of non invasive risk stratifiers assessing recurrent arrhythmias in ICD-recipients Parameter

Regression coefficient B

Exp (B)

P-value

Chemoreflex sensitivity ,3.0 ms / mmHg Baroreflex sensitivity ,3.0 ms / mmHg Late potential positive Heart rate variability ,30 ms LV ejection fraction ,40%

2.04 0.11 1.16 0.44 1.67

7.68 1.12 3.19 1.54 5.28

0.03 n.s. n.s. n.s. 0.04

174

M.G. Hennersdorf et al. / International Journal of Cardiology 86 (2002) 169–175

myocardial infarction [7,19,20] or with heart failure [21]. In 1998, La Rovere et al. [7] published the results of a large multicenter study (the ATRAMI trial) and showed that patients with an arrhythmic event after myocardial infarction are presenting with a reduced baroreflex sensitivity. The relative risk of a baroreflex sensitivity below 3 ms / mmHg amounted to 4.5 (95% CI 2.17–9.24). But at the present time, patients with an increased risk for ventricular tachyarrhythmias are not identifiable with sufficient accuracy using the conventional methods. As a new method for detection of disturbances of the vagal reflex activity the chemoreflex sensitivity is described in this study. The employed analysis of the chemoreflex sensitivity is very simple to perform [12]. By inhalation of pure oxygen, a deactivation of the arterial chemoreceptors results. This deactivation leads to a vagal stimulation and a decrease in heart rate [11,22–24]. But this decrease in the heart rate is dependent on a disturbed vagal reflex activity. Patients with survived ventricular arrhythmias show a significantly decreased chemoreflex sensitivity [12]. As previous studies showed, the sensitivity and specificity of chemoreflex sensitivity amount to 81 and 89%, respectively regarding a survived tachyarrhythmic event [25]. The present study reports the results of a prospective study for the evaluation of the predictive accuracy of the chemoreflex sensitivity in rhythmologically high risk patients. The Kaplan– Meier curves for arrhythmia free survival were significantly different for patients with and without depressed chemoreflex sensitivity. The positive predictive value regarding an arrhythmic event amounted to 61%, and the relative risk to 2.83 (95% CI 0.99– 8.01). Another independent risk factor for arrhythmia recurrence was a left ventricular function below 40%. Other known risk factors, such as baroreflex sensitivity or heart rate variability, did not reach the level of significance in this study. Hohnloser et al. [26] also could not find a significant difference in patients with or without decreased baroreflex sensitivity in a similar patient cohort. Other studies investigating the chemoreflex describe the analysis of the arterial oxygen pressure [22,24]. We used a method in this study that ought to be very simple by analyzing the venous partial oxygen pressure. During oxygen inhalation both values, arterial and venous oxygen pressure, can be

elevated. Previous studies [12] showed that the chemoreflex sensitivities, determined by arterial or venous oxygen pressure, are correlating well. Hence, the venous oxygen pressure can be used for the analysis of chemoreflex sensitivity. The possible reasons for a decreased chemoreflex sensitivity have to be hypothesized. Changes in the sensitivity of the arterial chemoreceptors, a changed behaviour of the receptors in reacting on catecholamines and central disorders have to be discussed. Interferences with other reflexes, i.e. the baroreflex, are reported. Somers et al. [27] showed an inhibitory interaction of the baroreflex and the peripheral chemoreflex and explained this by the convergence of baroreceptor and peripheral chemoreceptor afferents on neurons in the medulla. Probably, the chemoreflex sensitivity and the baroreflex sensitivity are of different impact in predicting life threatening ventricular arrhythmias or sudden cardiac death. The afferents are stimulated in different ways (pO 2 , respectively arterial pressure). Correlations between chemoreflex sensitivity and hemodynamic parameters were not found, whereas the baroreflex is correlating with the stroke volume index in patients with heart failure [28]. Additionally, there are different localizations in the central nervous system of the neurons of baroreceptors and chemoreceptors [29], and the age dependency is different between these two reflexes [30]. Thus, different information about vagal reflex activity could be expected from baroreflex sensitivity and chemoreflex sensitivity.

4.1. Limitations The analysis of the heart rate variability was performed over a short period of 15 min. This is different from the cited publications. It might be possible that the long period analysis of heart rate variability over 24 h could have led to different results. The patient cohort consists of individuals with survived ventricular tachycardia or ventricular fibrillation. This represents a population with a high risk of recurrent arrhythmias itself and constitutes a limitation of the study. But the goal of the study was to evaluate prospectively the potential impact of a new non invasive diagnostic test. In conclusion, this study shows a high predictive accuracy of the chemoreflex sensitivity in patients

M.G. Hennersdorf et al. / International Journal of Cardiology 86 (2002) 169–175

with implanted ICD and a previous survived myocardial infarction concerning future events. Further studies in larger trials especially in combination with other non invasive risk stratifiers, such as the t-wave alternans analysis, without arrhythmic event in their prior history are necessary to corroborate this investigation.

References [1] Vester EG, Strauer BE. Ventricular late potentials: state of the art and future perspectives. Eur Heart J 1994;15(Suppl C):34–48. [2] Simson MB. Use of signals in the terminal QRS-complex to identify patients with ventricular tachycardia after myocardial infarction. Circulation 1981;64:235–42. [3] Rosenbaum DS, Jackson LE, Smith JM, Garan H, Ruskin JN, Cohen RJ. Electrical alternans and vulnerability to ventricular arrhythmias. N Engl J Med 1994;330:235–41. [4] Glancy JM, Garratt CJ, Woods KL, de Bono DP. QT dispersion and mortality after myocardial infarction. Lancet 1995;345:945–8. [5] the Multicenter Post-Infarction Research Group, Kleiger RE, Miller JP, Bigger JTJ, Moss AJ. Decreased heart rate variability and its association with increased mortality after myocardial infarction. Am J Cardiol 1987;59:256–62. [6] Algra A, Tijssen JGP, Roelandt JRTC, Pool J, Lubsen J. Heart rate variability from 24-hour electrocardiography and the 2-year risk for sudden cardiac death. Circulation 1993;88:180–5. [7] for the ATRAMI (Autonomic Tone and Reflexes After Myocardial Infarction) Investigators, La Rovere MT, Bigger JTJ, Marcus FI, Mortara A, Schwartz PJ. Baroreflex sensitivity and heart-rate variability in prediction of total cardiac mortality after myocardial infarction. Lancet 1998;351:478–84. [8] Hohnloser S, Klingenheben T, van de Loo A, Hablawetz E, Just H, Schwartz PJ. Reflex versus tonic vagal activity as a prognostic parameter in patients with sustained ventricular tachycardia or ventricular fibrillation. Circulation 1994;89:1068–73. [9] Farrell TG, Odemuyiwa O, Bashir Y et al. Prognostic value of baroreflex sensitivity testing after myocardial infarction. Br Heart J 1992;67:129–37. [10] Ponikowski P, Chua TP, Piepoli M et al. Augmented peripheral chemosensitivity as a potential input to baroreflex impairment and autonomic imbalance in chronic heart failure. Circulation 1997;96:2586–94. [11] Groh WJ, Shinn TS, Engelstein EE, Zipes DP. Amiodarone reduces the prevalence of T wave alternans in a population with ventricular tachyarrhythmias. J Cardiovasc Electrophysiol 1999;10:1335–9. [12] Hennersdorf M, Perings C, Kristovic M et al. Chemoreflex- and baroreflexsensitivity among patients with survived sudden cardiac death. Z Kardiol 1997;86:196–203. [13] Dodge HT, Sandler H, Ballew DH, Lord JD. The use of biplane angiocardiography for the measurement of left ventricular volume in man. Am Heart J 1960;60:776–80. [14] Grassi G, Seravalle G, Cattaneo BM et al. Sympathetic activation and loss of reflex sympathetic control in mild congestive heart failure. Circulation 1995;92:3206–11.

175

[15] La Rovere MT, Specchia G, Mortara A, Schwartz PJ. Baroreflex sensitivity, clinical correlates, and cardiovascular mortality among patients with a first myocardial infarction. A prospective study. Circulation 1988;78:816–24. [16] Vester EG, Emschermann C, Stobbe U et al. Late potentials and heart rate variability in heart muscle disease. Eur Heart J 1994;15(Suppl C):25–33. [17] Nolan J, Batin PD, Andrews R et al. Prospective study of heart rate variability and mortality in chronic heart failure. Results of the United Kingdom Heart Failure Evaluation and Assessment of Risk Trial (UK-HEART). Circulation 1998;98:1510–6. [18] Eckberg DL, Drabinski M, Braunwald E. Defective cardiac parasympathetic control in patients with heart disease. N Engl J Med 1971;285:877–84. [19] Schwartz PJ, Zaza A, Pala M, Locati E, Beria G, Zanchetti A. Baroreflex sensitivity and its evolution during the first year after myocardial infarction. J Am Coll Cardiol 1988;12:629–36. [20] Schwartz PJ, La Rovere MT, Vanoli E. Autonomic nervous system and sudden cardiac death. Experimental basis and clinical observations for post-infarction risk stratification. Circulation 1992;85(Suppl I):I-77–I-91. [21] Osterziel KJ, Hanlein D, Willenbrock R, Eichhorn C, Luft F, Dietz R. Baroreflex sensitivity and cardiovascular mortality in patients with mild to moderate heart failure. Br Heart J 1995;73:517–22. [22] van de Borne PJ, Oren R, Anderson EA, Mark AL, Somers VK. Tonic chemoreflex activation does not contribute to elevated muscle sympathetic nerve activity in heart failure. Circulation 1996;94:1325–8. [23] Haque W, Boehmer J, Clemson BS, Leuenberger UA, Silber DH, Sinoway LI. Hemodynamic effects of supplemental oxygen administration in congestive heart failure. J Am Coll Cardiol 1996;27:353– 7. [24] Chua TP, Ponikowski P, Webb-Peploe K et al. Clinical characteristics of chronic heart failure patients with an augmented peripheral chemoreflex. Eur Heart J 1997;18:480–6. [25] Hennersdorf M, Perings C, Niebch V, Hillebrand S, Vester EG, Strauer BE. Chemoreflex sensitivity in patients with survived sudden cardiac arrest and prior myocardial infarction. Pacing Clin Electrophysiol 2000;23:457–62. [26] Hohnloser SH, Klingenheben T, Li YG, Zabel M, Peetermans J, Cohen RJ. T wave alternans as a predictor of recurrent ventricular tachyarrhythmias in ICD recipients: prospective comparison with conventional risk markers. J Cardiovasc Electrophysiol 1998;9:1258–68. [27] Somers VK, Mark AL, Abboud FM. Interaction of baroreceptor and chemoreceptor reflex control of sympathetic nerve activity in normal humans. J Clin Invest 1991;87:1953–7. ¨ [28] Osterziel K, Dietz R, Schmid W, Mikulschek K, Manthey J, Kubler W. ACE inhibition improves vagal reactivity in patients with heart failure. Am Heart J 1990;120:1120–9. [29] Chitravanshi VC, Sapru HN. Chemoreceptor-sensitive neurons in commissural subnucleus of nucleus tractus solitarius of the rat. Am J Physiol Regul Integr Comp Physiol 1995;268:R851–8. [30] Franchini KG, Moreira ED, Ida F, Krieger FM. Alterations in the cardiovascular control by the chemoreflex and the baroreflex in old rats. Am J Physiol Regul Integr Comp Physiol 1996;270:R310–3.