This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Email this article to a colleague Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Permissions Citing Articles Citing Articles via HighWire Citing Articles via Web of Science (1) Citing Articles via Google Scholar Google Scholar Articles by Chamchad, D. Articles by Horrow, J. C. Search for Related Content PubMed PubMed Citation Articles by Chamchad, D. Articles by Horrow, J. C. Related Collections Cardiovascular Complications Monitoring (Cardiac)

---

CARDIOVASCULAR ANESTHESIA

Nonlinear Heart Rate Variability Analysis May Predict Atrial Fibrillation After Coronary Artery Bypass Grafting

Dmitri Chamchad, MD^*, George Djaiani, MD, Hyun Ju Jung, MD, Lev Nakhamchik, MSc, Jo Carroll, RN, and Jay C. Horrow, MD, MS

From the *Department of Anesthesia, Lankenau Hospital, Wynnewood, Pennsylvania; Department of Anesthesia, Toronto General Hospital, Toronto, Ontario, Canada; and Department of Anesthesia, Drexel University College of Medicine, Philadelphia, Pennsylvania.

Address correspondence and reprint requests to Jay Horrow, MD, MS, Mail Stop 310, Broad and Vine, Philadelphia, PA 19102-1192. Address e-mail to jhorrow{at}drexelmed.edu.

	Abstract

Top Abstract Introduction METHODS RESULTS DISCUSSION REFERENCES

 
BACKGROUND: Heart rate variability might predict arrhythmias after coronary artery bypass grafting.

METHODS: Off-line processing of 10-min electrocardiogram recordings of consecutive patients provided R–R intervals for time domain, frequency domain, Poincaré, and point correlation analyses and subsequent association with postoperative atrial fibrillation by stepwise multivariate logistic regression.

RESULTS: Of 88 patients who met entry criteria, 13 developed atrial fibrillation. Peak point correlation dimension (odds ratio 3.985/unit, P = 0.0096) and age (odds ratio 1.144/yr, P = 0.0019) were independently associated with atrial fibrillation (c-statistic = 0.839).

CONCLUSIONS: Further study should confirm the ability of peak point correlation dimension to predict atrial fibrillation after coronary artery surgery with cardiopulmonary bypass.

	Introduction

Top Abstract Introduction METHODS RESULTS DISCUSSION REFERENCES

 
Postoperative cardiac arrhythmias occur in 15%–45% of cardiac surgery patients (1–8). Preoperative identification of patients at risk for postoperative atrial fibrillation might permit targeted prophylaxis. We evaluated the associations of traditional and newer, nonlinear, analytic heart rate variability (HRV) techniques with atrial fibrillation after coronary artery bypass grafting (CABG) with cardiopulmonary bypass (CPB).

	METHODS

Top Abstract Introduction METHODS RESULTS DISCUSSION REFERENCES

 
The protocol was approved by the ethics committee of Toronto General Hospital; patients provided oral consent per ethics committee directive. Eligible patients were those at least 18 yr of age of either sex scheduled for elective CABG with CPB. Patients who had a history of heart surgery, atrial fibrillation or flutter, sick sinus syndrome, or a cardiac pacemaker were excluded.

Aside from oral lorazepam 2–4 mg premedication, no other medications were given to patients before initial electrocardiogram (ECG) recording. Anesthesia consisted of IV propofol, fentanyl, or sufentanil, and volatile anesthetic. All patients underwent continuous ECG telemetry monitoring with automatic arrhythmia detection software for 5 days after surgery, recording all atrial and ventricular arrhythmias. Atrial fibrillation was defined as at least 24 h of an irregularly-irregular pattern of QRS-waves on the displayed and printed ECG.

Each patient underwent a 10-min digitally stored ECG recording within 2 h of scheduled surgery. Off-line analysis of the RR intervals used R-peak detection software (Windaq, Dataq Instruments, Akron, OH). Various analyses of the beat-to-beat variations in the RR intervals generated the HRV results in the time and frequency domains (9–11). For time domain analysis, we calculated the mean and standard deviation of RR intervals, the root mean square of successive RR interval differences (rMSSD), the number of RR intervals for which successive RR intervals differed by at least 50 ms (NN50), and the number of times successive RR intervals differed by >50% from the index RR interval (pNN50). The integral of the RR density distribution divided by the maximum of the density distribution (RR triangular), and the length of the base of a triangle approximating the NN interval distribution (TINN) were also determined. Frequency domain analyses included the following: power spectral density using a recursive maximum entropy method (MEM-all poles); low-frequency (LF: 0.04–0.15 Hz) and high-frequency (HF: 0.15–0.40 Hz) components using proprietary, nonparametric, fast Fourier transform software (Biomedical Signal Analysis Group, University of Kuoplo, Finland); percentages of total power in the LF and HF ranges; and the ratio of LF to HF power.

SAS software version 2.5D (SAS Institute, Cary, NC) produced a Poincaré plot of each RR interval against the previous RR interval (Fig. 1) and for all RR intervals, calculated dispersions of the short-term (SD1) and continuous long-term (SD2) intervals and their ratio (12,13). Point correlation dimension (PD2) reflects the complexity of information generated. Similar to a fractal, the PD2 is not restricted to a whole number, and represents the magnitude of independent sources of variability in a signal. Both mean and peak PD2 were calculated for each patient using Chaos Software (Bangor, PA).

View larger version (9K): [in this window] [in a new window]   Figure 1. Poincaré plot representing two patients with different types of heart rate variability. Each RR interval (RRj) in milliseconds is plotted against the immediately following one (RRj + 1). The dark points, in cloud-like formation, represent the data from a patient with a high HRV index, whereas the light points, in linear formation, come from a patient with low HRV index. Note the disparate values of peak point correlation dimension (pPD2) for each of these patients.

Atrial arrhythmia was the primary outcome for determination of sample size. Potential predictor variables with a univariate P < 0.05 by Wald statistics underwent stepwise multivariate logistic regression using SAS software version 9 (SAS Institute). The c-statistic, equivalent to the area under the receiver operating characteristic (ROC) curve, evaluated overall significance of the regression; a c-statistic of 0.5 represents poor predictive ability; values >0.60, a good predictive ability, and a value of 1.0, a perfect ability. Demographic, clinical, and HRV data are reported as mean ± sd unless otherwise noted.

	RESULTS

Top Abstract Introduction METHODS RESULTS DISCUSSION REFERENCES

 
Of 120 screened patients, 22 were excluded for preexisting arrhythmias and 10 for planned surgery without CPB ("off-pump"). Of 88 enrolled patients, 17 (19%) developed atrial fibrillation after surgery. Compared with patients remaining in sinus rhythm, those with atrial fibrillation were older and more likely to have hypertension at, but not before, CABG (Table 1).

Results for the HRV analysis for patients with and without postoperative atrial fibrillation are listed in Table 2. The following preoperative HRV results were significantly increased in patients who developed atrial fibrillation postoperatively compared with those remaining in sinus rhythm: the percentage of total power in the HF range, peak PD2, and mean PD2.

Table 3 shows results of the multivariate logistic regression. After adjusting for demographic and other differences between the groups, both patient age and peak PD2 were independently associated with postoperative atrial fibrillation (c-statistic of 0.839; see Table 3 for odds ratios and their confidence intervals).

	DISCUSSION

Top Abstract Introduction METHODS RESULTS DISCUSSION REFERENCES

 
These data demonstrate an independent association between patient age and peak point correlation dimension of HRV (pPD2) measured preoperatively with the development of atrial fibrillation after CABG surgery.

Postoperative arrhythmia is common after cardiac surgery (1–8). In the current study, all arrhythmias occurred in 38.6% of patients after CABG surgery with CPB. Previous studies with larger cohorts report an incidence of between 15% and 40%, and increased incidence with increasing age. The current study confirmed that age associates strongly with atrial fibrillation after CABG surgery with CPB.

Although many studies attempt to predict postoperative cardiac arrhythmias, the only consistent predictor among them is advanced age (2,3,5–8,12,14–18). In the current trial, the frequency domain measure of LF power, which tracks sympathetic activity, and the linear measures pNN50 and rMSSD, which track vagal activity, did not predict patients who developed atrial fibrillation postoperatively. Of the linear, frequency domain, quantitative Poincaré, and PD2 measurements of HRV in this trial, only pPD2 survived the multivariate regression process. The odds ratio for pPD2, nearly fourfold per 1.0 dimensional unit, demonstrates that pPD2 is a potent associative factor for postoperative atrial fibrillation, considering the small variability of pPD2 in the current cohorts, i.e., standard deviations of 0.66 dimensional units for patients without and 0.91 units for those with postoperative atrial fibrillation.

Nonlinear analysis identifies the functional order and temporal unfolding of heart rate dynamics, which have many of the features of complex adaptive systems (11). PD2 views the heart rate-generating system as an information source and quantifies the number of its independent regulating inputs. This approach can discriminate better than conventional techniques for high-risk, abrupt, autonomic changes preceding acute coronary events, e.g., sudden cardiac death.

In the current study of patients undergoing CABG, larger pPD2 values were associated with atrial fibrillation. This result fits with current understanding of its physiology: less complex pacemaker activity could predispose to hypotension with decreased autonomic tone, whereas more complex pacemaker activity could predispose to tachy-arrhythmia with abrupt autonomic changes. A numerically small pPD2 represents a heart that works like a "metronome" without modulation by the autonomic nervous system. This yields a more stable environment and thus explains the association of postoperative atrial fibrillation with increasing pPD2.

This study has several limitations. It provided a small number of events for many potential factors, risking spurious associations and possibly missing subtle associations that a richer dataset would reveal (19). The emerging field of HRV still lacks standardization of data acquisition, limiting the value of cross-study, but not within-study, comparisons. These preliminary data need confirmation in a larger independent cohort to identify and apply a value of pPD2 for predicting atrial fibrillation. They address only CABG patients undergoing CPB. Predictors may differ for arrhythmias after off-pump revascularization or other cardiac operations.

Should pPD2 prove a robust predictor of postoperative arrhythmia in a subsequent validation cohort, it could identify which patients should receive prophylactic therapy to prevent this burdensome postoperative complication.

	Footnotes

 
Accepted for publication July 19, 2006.

Presented in part at the Annual Meeting of the American Society of Anesthesiologists, October 2004, Las Vegas, NV.

	REFERENCES

Top Abstract Introduction METHODS RESULTS DISCUSSION REFERENCES

 

1. Akyurek O, Diker E, Guldal M, Oral D. Predictive value of heart rate variability for the recurrence of chronic atrial fibrillation after electrical cardioversion. Clin Cardiol 2003;26:196–200.[Web of Science][Medline]
2. Ascione R, Reeves BC, Santo K, et al. Predictors of new malignant ventricular arrhythmias after coronary surgery: a case–control study. J Am Coll Cardiol 2004;43:1630–8.[Abstract/Free Full Text]
3. Copie X, Lamaison D, Salvador M, et al. Heart rate variability before ventricular arrhythmias in patients with coronary artery disease and an implantable cardioverter defibrillator. Ann Noninvasive Electrocardiol 2003;8:179–84.[Web of Science][Medline]
4. Crystal E, Connolly SJ, Sleik K, et al. Interventions on prevention of postoperative atrial fibrillation in patients undergoing heart surgery: a meta-analysis. Circulation 2002;106:75–80.[Abstract/Free Full Text]
5. Guo Y, Hu S, Wu Q, et al. Predictors of atrial fibrillation after coronary artery bypass graft surgery. Chin Med J (Engl) 2002;115:232–4.[Medline]
6. Mathew JP, Fontes ML, Tudor IC, et al. A multicenter risk index for atrial fibrillation after cardiac surgery. JAMA 2004;291:1720–9.[Abstract/Free Full Text]
7. Taylor AD, Groen JG, Thorn SL, et al. New insights into onset mechanisms of atrial fibrillation and flutter after coronary artery bypass graft surgery. Heart 2002;88:499–504.[Abstract/Free Full Text]
8. Yeung-Lai-Wah JA, Qi A, McNeill E, et al. New-onset sustained ventricular tachycardia and fibrillation early after cardiac operations. Ann Thorac Surg 2004;77:2083–8.[Abstract/Free Full Text]
9. Task Force of the European Society of Cardiology and the North American Society of Pacing and Electrophysiology. Heart rate variability: standards of measurement, physiological interpretation and clinical use. Circulation 1996;93:1043–65.[Free Full Text]
10. Goldberger AL, West BJ. Applications of nonlinear dynamics to clinical cardiology. Ann N Y Acad Sci 1987;504:195–213.[Web of Science][Medline]
11. Kresh JY, Izrailtyan I. Evolution in functional complexity of heart rate dynamics: a measure of cardiac allograft adaptability. Am J Physiol 1998;275:R720–7.[Medline]
12. Hogue CW Jr, Domitrovich PP, Stein PK, et al. RR interval dynamics before atrial fibrillation in patients after coronary artery bypass graft surgery. Circulation 1998;98:429–34.[Abstract/Free Full Text]
13. Huikuri HV, Poutiainen AM, Makikallio TH, et al. Dynamic behavior and autonomic regulation of ectopic atrial pacemakers. Circulation 1999;100:1416–22.[Abstract/Free Full Text]
14. Bilchick KC, Fetics B, Djoukeng R, et al. Prognostic value of heart rate variability in chronic congestive heart failure (Veterans Affairs' Survival Trial of Antiarrhythmic Therapy in Congestive Heart Failure). Am J Cardiol 2002;90:24–8.[Web of Science][Medline]
15. Aranki SF, Shaw DP, Adams DH, et al. Predictors of atrial fibrillation after coronary artery surgery. Current trends and impact on hospital resources. Circulation 1996;94:390–7.[Abstract/Free Full Text]
16. Brouwer J, van Veldhuisen DJ, Man in 't Veld AJ, et al; the Dutch Ibopamine Multicenter Trial Study Group. Prognostic value of heart rate variability during long-term follow-up in patients with mild to moderate heart failure. J Am Coll Cardiol 1996;28:1183–9.[Abstract]
17. Hakala T, Vanninen E, Hedman A, Hippelainen M. Analysis of heart rate variability does not identify the patients at risk of atrial fibrillation after coronary artery bypass grafting. Scand Cardiovasc J 2002;36:167–71.[Web of Science][Medline]
18. Hopf HB, Skyschally A, Heusch G, Peters J. Low-frequency spectral power of heart rate variability is not a specific marker of cardiac sympathetic modulation. Anesthesiology 1995;82:609–19.[Web of Science][Medline]
19. Peduzzi P, Concato J, Kemper E, et al. A simulation study of the number of events per variable in logistic regression analysis. J Clin Epidemiol 1996;49:1373–9.[Web of Science][Medline]

This article has been cited by other articles:

				C. Taneyama and H. Goto Fractal Cardiovascular Dynamics and Baroreflex Sensitivity After Stellate Ganglion Block Anesth. Analg., October 1, 2009; 109(4): 1335 - 1340. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Email this article to a colleague Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Permissions Citing Articles Citing Articles via HighWire Citing Articles via Web of Science (1) Citing Articles via Google Scholar Google Scholar Articles by Chamchad, D. Articles by Horrow, J. C. Search for Related Content PubMed PubMed Citation Articles by Chamchad, D. Articles by Horrow, J. C. Related Collections Cardiovascular Complications Monitoring (Cardiac)