- Email: [email protected]
The relative importance of left atrial function versus dimension in predicting atrial fibrillation after coronary artery bypass graft surgery
The relative importance of left atrial function versus dimension in predicting atrial fibrillation after coronary artery bypass graft surgery Toshiko Nakai, MD,a Randall J. Lee, MD, PhD,a Nelson B. Schiller, MD,b Wayne H. Bellows, MD,d Samir Dzankic, MD,c John Reeves III, MD,d Joseph Romson, MD, PhD,d Scott Ferguson, MD,f and Jacqueline M. Leung, MD, MPHc San Francisco, Calif
Background Atrial fibrillation (AF) is a common complication after coronary artery bypass graft (CABG) surgery. The purpose of this study was to determine whether pre-existing left atrial dysfunction is a predictor of postoperative AF compared with other clinical predictors.
Methods Ninety-three patients undergoing CABG were prospectively studied. Intraoperatively, transesophageal echocardiography was performed to measure left atrial size, transmitral flow velocity, and other routine parameters. Left atrial function was estimated by the following formula: Atrial index = Transmitral VTI total × LAEF/Left atrial maximal area (where VTI = velocity time integral of E and A waves, LAEF = left atrial ejection fraction). The association of potential clinical predictors with the occurrence of postoperative AF was evaluated by χ2 or Fisher exact tests, followed by stepwise multivariate logistic regression model. P values and odds ratios (OR) with 95% CIs were reported. Significance was set at P < .05.
Results Postoperative AF occurred in 28 of 93 patients (30.1%). Patients with postoperative AF were older (67.0 ± 8.3 vs 61.5 ± 9.6 years, P = .0075), had larger left atrial maximal area (14.3 ± 4.6 cm2 vs 10.9 ± 4.3 cm2, P < .001), lower atrial index (0.54 ± 0.56 vs 0.82 ± 0.64, P = .008), larger body surface area (BSA) (OR 57, 95% CI 3.97-827), longer aortic cross-clamp time (OR 1.03, 95% CI 1.00-1.05), and more likely to have a postoperative myocardial infarction (OR 3.28, 95% CI 0.99-10.87) compared with those without AF. By multivariate analysis, only age (OR 1.11, 95% CI 1.04–1.19, P = .002) and atrial dimension (OR 1.75, 95% CI 1.03–2.96, P = .038) were significant independent predictors of postoperative AF. Body surface area also increased the odds of postoperative AF, but the CI was wide (OR 114, 95% CI 4.65–2810, P = .004). Conclusions Our results demonstrate that age and atrial enlargement, rather than atrial function, were independent predictors of postoperative AF. (Am Heart J 2002;143:181-6.) Atrial fibrillation (AF) is a common complication after coronary artery bypass graft (CABG) surgery, occurring in 20% to 40% of patients.1-3 Postoperative AF may lead to compromised heart function, stroke, and thromboembolism, iatrogenic complications with accompanying longer and more costly hospitalizations.3-6 Despite the identification of potential predictors and trials of prophylactic therapy,1-4,7-14 no consistent therapies have been From the aDepartment of Medicine, Section of Cardiac Electrophysiology, Department of Medicine and Cardiovascular Research Institute, and Departments of bMedicine and cAnesthesia and Perioperative Care, University of California, San Francisco, and the Departments of dCardiovascular Anesthesia and eIntensive Care, Kaiser Permanente Medical Center, San Francisco, Calif. Supported in part by institutional funds and the National Institute of Aging, National Institutes of Health grant No. 1K24 AG00948-01A2 (J. M. L.). Submitted June 20, 2001; accepted September 18, 2001. Reprint requests: Toshiko Nakai, MD, University of California, San Francisco, Division of Cardiology, Cardiac Electrophysiology, 500 Parnassus Ave, MU-434, Box 1354, San Francisco, CA 94143. E-mail: [email protected] Copyright © 2002 by Mosby, Inc. 0002-8703/2002/$35.00 + 0 4/1/120294 doi:10.1067/mhj.2002.120294
found to be completely effective in preventing postoperative AF. One of the reasons the current therapy has been ineffective in preventing the occurrence of AF is that there is a lack of understanding of the pathophysiologic mechanisms of AF after open-heart surgery. In previous studies, the role of left atrial function as a predictor of AF after CABG has not been studied. Accordingly, we designed a study to test the hypothesis that AF is the result of atrial failure, which may be detected sensitively through the use of echocardiographic Doppler measurement of atrial and mitral inflow flow properties.
Methods Study patients Approval was obtained from the institutional Committee on Human Research. The study was exempted from the requirement of obtaining informed consent from patients because it involved the collection of existing records or diagnostic data that were publicly available. Inclusion criteria include patients undergoing elective CABG without concomitant cardiac or noncardiac surgical procedures. Exclusion criteria include
American Heart Journal January 2002
182 Nakai et al
patients in current AF or who have recently terminated AF (within 1 week of the current planned procedure), an uninterpretable echocardiographic image for the assessment of atrial and ventricular function, and patients with esophageal disease precluding the insertion of the transesophageal echocardiographic probe or preoperative atrial or ventricular pacemaker.
Intraoperative management Standard monitors for patients undergoing CABG surgery were used. These included 5-lead electrocardiography, radial and pulmonary arterial catheters, and transesophageal echocardiography (TEE). Intraoperative TEE was routinely performed, and monitoring was continuous throughout the intraoperative period in patients undergoing open heart surgery at our institutions. Anesthetic and surgical management of patients was not controlled and was per usual clinical practice. Cardiopulmonary bypass was conducted by use of a membrane oxygenator with hemodilution and mild systemic hypothermia.
TEE All patients had serial intraoperative TEE performed at 2 different time points during surgery: before sternotomy and after chest closure. All offline echocardiographic analysis was performed with the measurement programs from the Tomtec/Freeland System (Broomfield, Colo). All analyses were based on the average obtained from measurements from 2 beats. Mitral inflow velocities were measured at end expiration at the level of the 4-chamber view. The E- and A-wave velocities (mitral inflow), duration, E/A ratio, and the average deceleration time of the E wave were measured. Left atrial (LA) area is the average planimetered tracing of the endocardium and the mitral annulus of the 4-chamber (0degree transducer rotation) and the 2-chamber (90-degree transducer rotation) views. LA length is the average equidistant tracing from the mitral annulus to the LA posterior wall of the 4-chamber and 2-chamber views. If the atrium was foreshortened by the echocardiography field, the boundaries of the field were traced. LA ejection fraction (LAEF) was calculated to estimate LA reservoir function: LAEF = (LA maximal area – LA minimal area)/LA maximal area × 100%.
Atrial index In addition to estimating LA function by use of LAEF and LA appendage peak emptying velocity, we also used an “atrial index” developed previously by our group to estimate atrial function.15 The atrial index combines expressions of thefunctional milieu (ie, stoke distance) of the atrium, its total reservoir function (LAEF), and its state of dilation (LA maximal area). In the current study, we have adopted this index in the analysis of atrial function with TEE data. Atrial index = Transmitral VTI × LAEF/LA maximal area
where VTI is velocity time integral of E and A waves.
P-wave duration and amplitude A preoperative baseline 12-lead electrocardiogram (ECG) was recorded in all patients. From lead II, P-wave duration was measured as the time interval between the onset of Pwave deflection to the end, P-wave amplitude was measured
as the absolute value from the baseline to positive or negative maximum deflection, and the biphasic P-wave amplitude was the sum of both the absolute value of positive and negative maximum deflections.
Outcome ascertainment All patients were followed up daily with continuous ECG telemetry for the occurrence of postoperative AF (episode of irregularly irregular rhythm requiring clinical intervention such as cardioversion or pharmacologic therapy or any episode lasting ≥10 minutes) during the in-hospital period. All outcomes were validated by 2 investigators. Postoperative myocardial infarction was defined as new Q waves (≥40 milliseconds, 25% R wave) on the postoperative 12-lead ECG and a creatine phosphokinase-MB isoenzyme concentration ≥50 U/L.
Statistical analysis Patients were divided into 2 groups on the basis of the development of AF. All data were presented as mean ± SD. Student t test or a rank sum test was used for statistical analysis. The association of potential clinical predictors with the occurrence of postoperative AF was evaluated with the χ2 or Fisher exact tests (Stata 5.0, Stata Corporation, College Station, Tex). Variables that had significant association with postoperative AF on univariate analysis (P value ≤.1) were entered in a stepwise multivariate logistic regression model (STATA statistical software version 5.0, Stata Corporation). P values and odds ratios (OR) with 95% CI were reported. P < .05 was considered statistically significant. Sensitivity and specificity of left atrial dimension in predicting postoperative AF were calculated by standard formulas.
Results Of 140 eligible patients, 47 patients were excluded as a result of inadequate echocardiographic images. Inadequate images included any of the missing echocardiographic parameters measured in either the prebypass or the postbypass measurement periods. The total study population in the current analysis consisted of 93 patients (74 men; aged 42-81 years, mean age 63.2 ± 9.5 years). The characteristics of the study patients are shown in Table I. Twenty-eight of 93 patients (30.1%) had postoperative AF. Patients who had AF were significantly older than those without AF (67.0 ± 8.3 vs 61.5 ± 9.6 years, P = .0075). The prevalence of postoperative AF increases with age: patients aged >60 years had a 2-fold higher rate of postoperative AF than those <60 years old (Figure 1). The type of preoperative cardiac medications had no association with postoperative AF. In this study, no prophylactic antiarrhythmic agents were used. Other clinical characteristics associated with postoperative AF include larger body surface area and a history of AF.
TEE findings TEE data, which were measured before sternotomy, are shown in Table II. Patients with postoperative AF
American Heart Journal Volume 143, Number 1
Nakai et al 183
Table I. Clinical characteristics of patients with and without AF after CABG
Age (y) Sex (male/female) Body surface area Habits Smoking Alcohol Disease History of AF History of myocardial infarction Hypertension Diabetes mellitus Medications β-Blocker Calcium channel blocker Angiotensin-converting enzyme inhibitor Cross-clamp time (min) Postoperative myocardial infarction Hospital stay after CABG (d)
non-AF (n = 65)
AF (n = 28)
P value
61.5 ± 9.6 49/16 1.92 ± 0.19
67.0 ± 8.3 25/3 2.06 ± 0.21
.008 .213 .001
39 21
17 13
.868 .288
0 23 42 25
4 10 17 10
.007 .837 .902 .986
46 22 21 54.3 ± 17.0 3 6.4 ± 9.6
18 5 6 64.1 ± 22.9 4 8.8 ± 7.6
.774 .190 .417 .114 .192 <.001
Data are expressed as mean value ± SD.
Table II. TEE data
LA maximal area (cm2) LAEF (%) Transmitral VTI E/A LA appendage emptying velocity (m/s) Atrial index LVEF (%)
Non-AF (n = 65)
AF (n = 28)
P value
10.9 ± 4.3 34.0 ± 10.5 1.98 ± 0.86 0.46 ± 0.15 0.82 ± 0.64 50.7 ± 12.6
14.3 ± 4.6 30.0± 10.9 1.86 ± 0.72 0.48 ± 0.21 0.54 ± 0.56 51.5 ± 11.8
<.001 .092 .492 .550 .008 .786
had larger atria as measured by LA maximal area than did those without postoperative AF (14.3 ± 4.6 cm2 vs 10.9 ± 4.3 cm2, P < .001). LA maximal area >10.5 cm2 had a sensitivity of 85.7% and a specificity of 58.5% in predicting postoperative AF. There was no significant difference in LAEF, transmitral VTI E/A, LA appendage emptying velocity, and left ventricular ejection fraction (LVEF) between patients with and without AF. The atrial index in the group with postoperative AF was significantly lower than in those without AF (0.54 ± 0.56 vs 0.82 ± 0.64, P = .008). The sensitivity and specificity of atrial index in (<0.64) predicting postoperative AF was 85.7% and 49.2%, respectively. Figure 2 shows the values of atrial index before sternotomy and after chest closure during the operation. There was a significant difference in the mean atrial index between the groups with and without postoperative AF before sternotomy. The mean LVEF increased by 9.5% immediately after bypass intraoperatively in the group without postoperative AF (51% ± 13% to 53% ± 12%) but remained unchanged (mean change 0.7%) in the group who subsequently had postoperative AF (52% ± 12% to 50 ± 11%). However, the between-group change in LVEF was not significantly different (9.5% vs 0.7%, P = .08). Also, the change in LVEF was not different between
Figure 1
Incidence of postoperative AF. The percentages of patients with postoperative AF for each of the age group (10-y intervals) are shown. See text for details.
patients who had postoperative myocardial infarction versus those without such complications (10.7% vs –1.3%, P = .37).
American Heart Journal January 2002
184 Nakai et al
Figure 2
Table III. Univariate predictors of postoperative AF Variables
OR (95% CI%)
P value
Age Atrial index LA dimension Body surface area Aortic cross-clamp time Postoperative myocardial infarction
1.07 (1.01-1.25) 0.38 (0.14-1.04) 2.10 (1.29-3.40) 57 (3.97-827) 1.03 (1.00-1.05) 3.28 (0.99-10.86)
.013 .059 .003 .003 .029 .052
Table IV. Multivariate analysis of postoperative AF Variables
Atrial index. The mean atrial indices for the patients with AF and without AF are shown. Pre, Prebypass period; post, postbypass period.
Age LA dimension BSA
OR (95% CI)
P value
1.11 (1.04-1.19) 1.74 (1.03-2.96) 114 (4.65-2810)
.002 .038 .004
Comparisons with previous studies ECG findings There were no significant differences in P-wave duration (112.6 ± 18.1 vs 114.5 ± 18.1 milliseconds, P = .89) or P-wave amplitude (1.09 ± 0.44 vs 1.09 ± 0.52 mV, P = .47) between the 2 groups.
Atrial index versus other clinical predictors By univariate analysis, increased age, lower atrial index, increased left atrial size, increased body surface area, longer aortic cross-clamp time, and postoperative myocardial infarction increased the odds of postoperative AF (Table III). All variables that were significant after univariate analysis were entered in a stepwise multivariate logistic regression model. In the final model, only age (OR 1.11, 95% CI 1.04–1.19, P = .002) and atrial dimension (OR 1.74, 95% CI 1.03–2.96, P = .038) were significant independent predictors of postoperative AF (Table IV). Body surface area also increased the odds of postoperative AF, but the CI was wide (OR 114, 95% CI 4.65–2810, P = .004), suggesting that larger sample size is necessary to confirm the accuracy of the estimate.
Discussion The major observations in our study are that advanced age and left atrial enlargement were independent predictors of AF after CABG. Although left atrial dysfunction was found to be a significant predictor of postoperative AF on univariate analysis, it did not remain as an independent predictor on multivariate analysis. P-wave analysis with standard 12-lead ECG did not provide predictive information of AF after CABG.
Our finding that 30.1% of patients had postoperative AF is similar to that reported by previous studies.1-3 Our results also agree with those from previous studies demonstrating that advanced age and LA enlargement were independent predictors of postoperative AF. In contrast, our study, along with others,6,16 did not confirm the finding that men have a higher risk of postoperative AF than women do.8,17 Whether sex is a surrogate of other predictors of postoperative AF remains to be determined.
LA dimension and function Although several studies had previously demonstrated that LA enlargement is predictive of AF after cardiac surgery, no previous study has investigated the role of atrial function in postoperative AF. Our study is the first to prospectively evaluate LA function by using intraoperative TEE as a predictor of AF after CABG. Our study was also the first to validate the utility of the LA index as a measurement of LA function. Our results demonstrated that the atrial index was significantly lower in the group who subsequently had postoperative AF than in those without postoperative AF, even before surgical intervention began, suggesting that pre-existist LA dysfunction may be a precursor for postoperative AF, independent of left ventricular function. After the operation, although atrial function was lower in patients who subsequently had postoperative AF compared with those without such complications, this difference did not reach statistical significance. Through atrial index and LA dimension analysis, it appears that the state of dilation of the atrium may be more important than the overall atrial function as represented by the total reservoir function (LAEF) and stroke distance (velocity time integral of the E and A waves) in predicting postoperative AF.
American Heart Journal Volume 143, Number 1
P-wave duration and P-wave amplitude Other studies found that prolonged P-wave duration was an independent predictor of AF after CABG.17 However, our study did not show significant difference in P-wave characteristics between groups with and without postoperative AF. Whether more detailed analysis such as signal-averaged P-wave duration will be of more predictive value will need to be determined by future studies.
Mechanism of postoperative AF The mechanism of postoperative AF is unclear. Many etiologic causes have been proposed, including inflammatory response causing atrial edema, pericarditis, or reperfusion injury8; atrial hypothermia resulting in prolonged conduction time and inducibility of AF because of shorter effective refractory period (ERP), dispersion of ERP, and slower conduction18; and hypoxic, ischemic, or hypokalemic changes resulting in shortening of the monophasic action potential duration.19 In addition, many studies also reported that advanced age is strongly associated with postoperative AF, as also demonstrated in the general nonsurgical population with AF.20 Our current results along with previous observations suggest that pre-existing age-related atrial remodeling, as evidenced by LA dilatation and decreased LA function, superimposed by the acute stress of surgical intervention, may have resulted in atrial ischemia and acute atrial failure, manifested clinically as postoperative AF.
Clinical relevance The current study demonstrated that LA dimension appeared to be more important than LA function in predicting postoperative AF. If LA function was relatively preserved in those who subsequently had postoperative AF, therapeutic regimens aimed at optimizing loading conditions of the atrium to preserve function may be important. These results suggest that conventional drugs such as angiotensin-converting enzyme inhibitors, which have been found beneficial in modulating ventricular remodeling in heart failure as well as chronic AF,21 may have an important impact in preventing AF through modulation of the atrial loading conditions. The use of preoperative or intrapreoperative echocardiographic measurement of LA dimension and function may have incremental value in identifying high-risk subjects, thus increasing the cost-effectiveness of prophylactic antiarrhythmic therapy.
Study limitations Our study focused only on the in-hospital period. In this study, the onset of the first AF episodes occurred predominantly in the early postoperative period (24-72 hours postoperatively), although episodes of AF can occur late in the in-hospital postoperative period. As a
Nakai et al 185
result, our study may have underestimated the incidence of postoperative AF.
Conclusions Advanced age and left atrial enlargement are independent predictors of AF after CABG. Our study also suggests that LA dimension appears to be more important than atrial function in predicting AF after CABG. Further study with a larger sample size will be necessary to confirm the independent predictive value of LA function. We thank Darwin Pastor and Kawalpreet Manku, MBBS, for assistance in data collection and analysis. We also thank the physicians in the Department of Cardiovascular Anesthesia and Department of Cardiovascular Surgery, Kaiser Permanente Medical Center, San Francisco, for their cooperation in making the study possible.
References 1. Creswell LL, Sschuessler RB, Rosenbloom M, et al. Hazards of postoperative atrial arrhythmias. Ann Thorac Surg 1993;56:539-49. 2. Mathew JP, Parks R, Savino JS, et al. Atrial fibrillation following coronary artery bypass graft surgery: predictors, outcomes, and resource utilization. Multicenter Study of Perioperative Ischemia Research Group. JAMA 1996;276:300-6. 3. Aranki SF, Shaw DP, Adams DH, et al. Predictors of atrial fibrillation after coronary artery surgery. Circulation 1996;94:390-7. 4. Taylor GJ, Malik SA, Colliver JA, et al. Usefulness of atrial fibrillation as a predictor of stroke after isolated coronary bypass grafting. Am J Cardiol 1987;60:905-7. 5. Reed GL III, Singer DE, Picard EH, et al. Stroke following coronary artery bypass surgery: a case-control estimate of the risk from carotid bruits. N Engl J Med 1988;319:1246-50. 6. Borzak S, Tisdale JE, Amin NB, et al. Atrial fibrillation after bypass surgery: does the arrhythmia or the characteristics of the patients prolong hospital stay? Chest 1998;113:1489-91. 7. Buxton AE, Josephson ME. The role of p wave duration as a predictor of postoperative atrial arrhythmias. Chest 1981;80:68-73. 8. Fuller JA, Adams GG, Buxton B. Atrial fibrillation after coronary artery bypass grafting. J Thorac Cardiovasc Surg 1989;97:821-5. 9. Steinberg JS, Zelenkofske S, Wong SC, et al. Value of the p-wave signal-averaged ECG for predicting atrial fibrillation after cardiac surgery. Circulation 1993;88:2618-22. 10. Mendes LA, Connelly GP, Mckenney PA, et al. Right coronary artery stenosis: an independent predictor of atrial fibrillation after coronary artery bypass surgery. J Am Coll Cardiol 1995;25:198202. 11. Zaman AG, Archbold AA, Helft G, et al. Atrial fibrillation after coronary bypass surgery: a model for preoperative risk stratification. Circulation 2000;101:1403-8. 12. Kempf FC, Hedberg A, Molinoff P, et al. The effect of pharmacologic therapy on atrial beta-receptor density and postoperative atrial arrhythmias [abstract]. Circulation 1983;68(III Suppl):57. 13. Ali IM, Sanalla AA, Clark V. Beta-blocker effects on postoperative atrial fibrillation. Eur J Cardiothorac Surg 1997;11:1154-7. 14. Frost L, Mortensen PE, Tingleff J, et al. Dofetilide Post-CABG Study Group: efficacy and safety of dofetilide, a new class III antiarrhyth-
American Heart Journal January 2002
186 Nakai et al
15.
16.
17.
18.
mic agent, in acute termination of atrial fibrillation or flutter after coronary bypass surgery. Int J Cardiol 1997;58:135-40. Hourigan LA, Schiller NB. The composite index of left atrial function: a noninvasive tool for quantitating left atrial function [abstract]. J Am Coll Cardiol 2000;35(A Suppl):458A. Crosby LH, Pifalo WB, Woll KR, et al. Risk factors for atrial fibrillation after coronary artery bypass grafting. Am J Cardiol 1990;66: 1520-2. Chang CM, Lee SH, Lu MJ, et al. The role of P wave in prediction of atrial fibrillation after coronary artery surgery. Int J Cardiol 1999; 68:303-8. Sato S, Yamaguchi S, Schuessler RB, et al. The effect of augmented
atrial hypothermia on atrial refractory period, condition, and atrial flutter/fibrillation in the canine heart. J Thorac Cardiovasc Surg 1992;104:297-306. 19. Pichlmaier AM, Lang V, Harringer W, et al. Prediction of the onset of atrial fibrillation after cardiac surgery using the monophasic action potential. Heart 1998;80:467-72. 20. Furberg CD, Psaty BM, Manolio TA, et al. Prevalence of atrial fibrillation in elderly subjects (the Cardiovascular Health Study). Am J Cardiol 1994;74:236-41. 21. Van Den Berg MP, Crijns HJGM, Van Veldhuisen DJ. Effects of lisinopril in patients with heart failure and chronic atrial fibrillation. J Card Fail 1995;1:355-64.
The following article is an AHJ Online Exclusive. Full text of this article is available at no charge at our Web site: www.mosby.com/ahj. Valvular and Congenital Heart Disease
Systemic platelet effects of contrast media: Implications for cardiologic research and clinical practice M. C. D. Dalby, MD,a S. J. Davidson, MD,b J. F. Burman, MD,b J. Clague, MD,a U. Sigwart, MD,a and S. W. Davies, MDa London, United Kingdom
Background Angiographic contrast media cause platelet activation and decrease aggregability in vitro. We have previously shown in vitro a significant antiplatelet effect of contrast media at the concentrations obtained locally in the coronary artery during angioplasty. It is not known, however, whether a systemic effect is present. Method Thirty patients undergoing diagnostic coronary angiography were prospectively randomized to receive the nonionic medium iohexol, ionic low-molecular-weight medium ioxaglate, or ionic highmolecular-weight medium diatrizoate. Platelet aggregability was measured before and after the investigation with whole blood electrical impedance aggregometry (WBEA) with collagen agonist and the PFA-100 (Dade, Miami, FL) platelet function analyzer with combined shear, collagen, and adenosine diphosphate as agonists. Results With WBEA, with iohexol no difference in impedance change was seen: (medians and ranges) before, 9.8 Ω (4.8-19.2 Ω) versus after, 9.6 Ω (2-19.2 Ω) (P not significant [NS]). With
ioxaglate a significant fall was seen: before, 8.6 Ω (6.4-15.2 Ω) versus after, 6.6 Ω (0-12.4 Ω) (P = .004). With diatrizoate a significant and greater fall was seen: before, 10.8 Ω (6.4-17.6 Ω) versus after, 6.6 Ω (0-10.8 Ω) (P = .002). With PFA, no difference in closure time was seen with any medium: iohexol before, 99 seconds (79-142 seconds) versus after, 142 seconds (63-128 seconds) (P NS); ioxaglate before, 120 seconds (75-258 seconds) versus after, 95 seconds (74258 seconds) (P NS); and diatrizoate before, 114.5 seconds (65250 seconds) versus after, 100.5 seconds (72-300 seconds) (P NS).
Conclusions Ionic but not nonionic contrast media have a systemic antiplatelet effect at diagnostic angiographic doses when measured with WBEA. Such an effect has not been shown before. This may explain the observed improved clinical outcome with ionic contrast media but also might confound platelet studies in coronary angioplasty. (Am Heart J 2002;143:e1.) 4/90/119998
Background Atrial fibrillation (AF) is a common complication after coronary artery bypass graft (CABG) surgery. The purpose of this study was to determine whether pre-existing left atrial dysfunction is a predictor of postoperative AF compared with other clinical predictors.
Methods Ninety-three patients undergoing CABG were prospectively studied. Intraoperatively, transesophageal echocardiography was performed to measure left atrial size, transmitral flow velocity, and other routine parameters. Left atrial function was estimated by the following formula: Atrial index = Transmitral VTI total × LAEF/Left atrial maximal area (where VTI = velocity time integral of E and A waves, LAEF = left atrial ejection fraction). The association of potential clinical predictors with the occurrence of postoperative AF was evaluated by χ2 or Fisher exact tests, followed by stepwise multivariate logistic regression model. P values and odds ratios (OR) with 95% CIs were reported. Significance was set at P < .05.
Results Postoperative AF occurred in 28 of 93 patients (30.1%). Patients with postoperative AF were older (67.0 ± 8.3 vs 61.5 ± 9.6 years, P = .0075), had larger left atrial maximal area (14.3 ± 4.6 cm2 vs 10.9 ± 4.3 cm2, P < .001), lower atrial index (0.54 ± 0.56 vs 0.82 ± 0.64, P = .008), larger body surface area (BSA) (OR 57, 95% CI 3.97-827), longer aortic cross-clamp time (OR 1.03, 95% CI 1.00-1.05), and more likely to have a postoperative myocardial infarction (OR 3.28, 95% CI 0.99-10.87) compared with those without AF. By multivariate analysis, only age (OR 1.11, 95% CI 1.04–1.19, P = .002) and atrial dimension (OR 1.75, 95% CI 1.03–2.96, P = .038) were significant independent predictors of postoperative AF. Body surface area also increased the odds of postoperative AF, but the CI was wide (OR 114, 95% CI 4.65–2810, P = .004). Conclusions Our results demonstrate that age and atrial enlargement, rather than atrial function, were independent predictors of postoperative AF. (Am Heart J 2002;143:181-6.) Atrial fibrillation (AF) is a common complication after coronary artery bypass graft (CABG) surgery, occurring in 20% to 40% of patients.1-3 Postoperative AF may lead to compromised heart function, stroke, and thromboembolism, iatrogenic complications with accompanying longer and more costly hospitalizations.3-6 Despite the identification of potential predictors and trials of prophylactic therapy,1-4,7-14 no consistent therapies have been From the aDepartment of Medicine, Section of Cardiac Electrophysiology, Department of Medicine and Cardiovascular Research Institute, and Departments of bMedicine and cAnesthesia and Perioperative Care, University of California, San Francisco, and the Departments of dCardiovascular Anesthesia and eIntensive Care, Kaiser Permanente Medical Center, San Francisco, Calif. Supported in part by institutional funds and the National Institute of Aging, National Institutes of Health grant No. 1K24 AG00948-01A2 (J. M. L.). Submitted June 20, 2001; accepted September 18, 2001. Reprint requests: Toshiko Nakai, MD, University of California, San Francisco, Division of Cardiology, Cardiac Electrophysiology, 500 Parnassus Ave, MU-434, Box 1354, San Francisco, CA 94143. E-mail: [email protected] Copyright © 2002 by Mosby, Inc. 0002-8703/2002/$35.00 + 0 4/1/120294 doi:10.1067/mhj.2002.120294
found to be completely effective in preventing postoperative AF. One of the reasons the current therapy has been ineffective in preventing the occurrence of AF is that there is a lack of understanding of the pathophysiologic mechanisms of AF after open-heart surgery. In previous studies, the role of left atrial function as a predictor of AF after CABG has not been studied. Accordingly, we designed a study to test the hypothesis that AF is the result of atrial failure, which may be detected sensitively through the use of echocardiographic Doppler measurement of atrial and mitral inflow flow properties.
Methods Study patients Approval was obtained from the institutional Committee on Human Research. The study was exempted from the requirement of obtaining informed consent from patients because it involved the collection of existing records or diagnostic data that were publicly available. Inclusion criteria include patients undergoing elective CABG without concomitant cardiac or noncardiac surgical procedures. Exclusion criteria include
American Heart Journal January 2002
182 Nakai et al
patients in current AF or who have recently terminated AF (within 1 week of the current planned procedure), an uninterpretable echocardiographic image for the assessment of atrial and ventricular function, and patients with esophageal disease precluding the insertion of the transesophageal echocardiographic probe or preoperative atrial or ventricular pacemaker.
Intraoperative management Standard monitors for patients undergoing CABG surgery were used. These included 5-lead electrocardiography, radial and pulmonary arterial catheters, and transesophageal echocardiography (TEE). Intraoperative TEE was routinely performed, and monitoring was continuous throughout the intraoperative period in patients undergoing open heart surgery at our institutions. Anesthetic and surgical management of patients was not controlled and was per usual clinical practice. Cardiopulmonary bypass was conducted by use of a membrane oxygenator with hemodilution and mild systemic hypothermia.
TEE All patients had serial intraoperative TEE performed at 2 different time points during surgery: before sternotomy and after chest closure. All offline echocardiographic analysis was performed with the measurement programs from the Tomtec/Freeland System (Broomfield, Colo). All analyses were based on the average obtained from measurements from 2 beats. Mitral inflow velocities were measured at end expiration at the level of the 4-chamber view. The E- and A-wave velocities (mitral inflow), duration, E/A ratio, and the average deceleration time of the E wave were measured. Left atrial (LA) area is the average planimetered tracing of the endocardium and the mitral annulus of the 4-chamber (0degree transducer rotation) and the 2-chamber (90-degree transducer rotation) views. LA length is the average equidistant tracing from the mitral annulus to the LA posterior wall of the 4-chamber and 2-chamber views. If the atrium was foreshortened by the echocardiography field, the boundaries of the field were traced. LA ejection fraction (LAEF) was calculated to estimate LA reservoir function: LAEF = (LA maximal area – LA minimal area)/LA maximal area × 100%.
Atrial index In addition to estimating LA function by use of LAEF and LA appendage peak emptying velocity, we also used an “atrial index” developed previously by our group to estimate atrial function.15 The atrial index combines expressions of thefunctional milieu (ie, stoke distance) of the atrium, its total reservoir function (LAEF), and its state of dilation (LA maximal area). In the current study, we have adopted this index in the analysis of atrial function with TEE data. Atrial index = Transmitral VTI × LAEF/LA maximal area
where VTI is velocity time integral of E and A waves.
P-wave duration and amplitude A preoperative baseline 12-lead electrocardiogram (ECG) was recorded in all patients. From lead II, P-wave duration was measured as the time interval between the onset of Pwave deflection to the end, P-wave amplitude was measured
as the absolute value from the baseline to positive or negative maximum deflection, and the biphasic P-wave amplitude was the sum of both the absolute value of positive and negative maximum deflections.
Outcome ascertainment All patients were followed up daily with continuous ECG telemetry for the occurrence of postoperative AF (episode of irregularly irregular rhythm requiring clinical intervention such as cardioversion or pharmacologic therapy or any episode lasting ≥10 minutes) during the in-hospital period. All outcomes were validated by 2 investigators. Postoperative myocardial infarction was defined as new Q waves (≥40 milliseconds, 25% R wave) on the postoperative 12-lead ECG and a creatine phosphokinase-MB isoenzyme concentration ≥50 U/L.
Statistical analysis Patients were divided into 2 groups on the basis of the development of AF. All data were presented as mean ± SD. Student t test or a rank sum test was used for statistical analysis. The association of potential clinical predictors with the occurrence of postoperative AF was evaluated with the χ2 or Fisher exact tests (Stata 5.0, Stata Corporation, College Station, Tex). Variables that had significant association with postoperative AF on univariate analysis (P value ≤.1) were entered in a stepwise multivariate logistic regression model (STATA statistical software version 5.0, Stata Corporation). P values and odds ratios (OR) with 95% CI were reported. P < .05 was considered statistically significant. Sensitivity and specificity of left atrial dimension in predicting postoperative AF were calculated by standard formulas.
Results Of 140 eligible patients, 47 patients were excluded as a result of inadequate echocardiographic images. Inadequate images included any of the missing echocardiographic parameters measured in either the prebypass or the postbypass measurement periods. The total study population in the current analysis consisted of 93 patients (74 men; aged 42-81 years, mean age 63.2 ± 9.5 years). The characteristics of the study patients are shown in Table I. Twenty-eight of 93 patients (30.1%) had postoperative AF. Patients who had AF were significantly older than those without AF (67.0 ± 8.3 vs 61.5 ± 9.6 years, P = .0075). The prevalence of postoperative AF increases with age: patients aged >60 years had a 2-fold higher rate of postoperative AF than those <60 years old (Figure 1). The type of preoperative cardiac medications had no association with postoperative AF. In this study, no prophylactic antiarrhythmic agents were used. Other clinical characteristics associated with postoperative AF include larger body surface area and a history of AF.
TEE findings TEE data, which were measured before sternotomy, are shown in Table II. Patients with postoperative AF
American Heart Journal Volume 143, Number 1
Nakai et al 183
Table I. Clinical characteristics of patients with and without AF after CABG
Age (y) Sex (male/female) Body surface area Habits Smoking Alcohol Disease History of AF History of myocardial infarction Hypertension Diabetes mellitus Medications β-Blocker Calcium channel blocker Angiotensin-converting enzyme inhibitor Cross-clamp time (min) Postoperative myocardial infarction Hospital stay after CABG (d)
non-AF (n = 65)
AF (n = 28)
P value
61.5 ± 9.6 49/16 1.92 ± 0.19
67.0 ± 8.3 25/3 2.06 ± 0.21
.008 .213 .001
39 21
17 13
.868 .288
0 23 42 25
4 10 17 10
.007 .837 .902 .986
46 22 21 54.3 ± 17.0 3 6.4 ± 9.6
18 5 6 64.1 ± 22.9 4 8.8 ± 7.6
.774 .190 .417 .114 .192 <.001
Data are expressed as mean value ± SD.
Table II. TEE data
LA maximal area (cm2) LAEF (%) Transmitral VTI E/A LA appendage emptying velocity (m/s) Atrial index LVEF (%)
Non-AF (n = 65)
AF (n = 28)
P value
10.9 ± 4.3 34.0 ± 10.5 1.98 ± 0.86 0.46 ± 0.15 0.82 ± 0.64 50.7 ± 12.6
14.3 ± 4.6 30.0± 10.9 1.86 ± 0.72 0.48 ± 0.21 0.54 ± 0.56 51.5 ± 11.8
<.001 .092 .492 .550 .008 .786
had larger atria as measured by LA maximal area than did those without postoperative AF (14.3 ± 4.6 cm2 vs 10.9 ± 4.3 cm2, P < .001). LA maximal area >10.5 cm2 had a sensitivity of 85.7% and a specificity of 58.5% in predicting postoperative AF. There was no significant difference in LAEF, transmitral VTI E/A, LA appendage emptying velocity, and left ventricular ejection fraction (LVEF) between patients with and without AF. The atrial index in the group with postoperative AF was significantly lower than in those without AF (0.54 ± 0.56 vs 0.82 ± 0.64, P = .008). The sensitivity and specificity of atrial index in (<0.64) predicting postoperative AF was 85.7% and 49.2%, respectively. Figure 2 shows the values of atrial index before sternotomy and after chest closure during the operation. There was a significant difference in the mean atrial index between the groups with and without postoperative AF before sternotomy. The mean LVEF increased by 9.5% immediately after bypass intraoperatively in the group without postoperative AF (51% ± 13% to 53% ± 12%) but remained unchanged (mean change 0.7%) in the group who subsequently had postoperative AF (52% ± 12% to 50 ± 11%). However, the between-group change in LVEF was not significantly different (9.5% vs 0.7%, P = .08). Also, the change in LVEF was not different between
Figure 1
Incidence of postoperative AF. The percentages of patients with postoperative AF for each of the age group (10-y intervals) are shown. See text for details.
patients who had postoperative myocardial infarction versus those without such complications (10.7% vs –1.3%, P = .37).
American Heart Journal January 2002
184 Nakai et al
Figure 2
Table III. Univariate predictors of postoperative AF Variables
OR (95% CI%)
P value
Age Atrial index LA dimension Body surface area Aortic cross-clamp time Postoperative myocardial infarction
1.07 (1.01-1.25) 0.38 (0.14-1.04) 2.10 (1.29-3.40) 57 (3.97-827) 1.03 (1.00-1.05) 3.28 (0.99-10.86)
.013 .059 .003 .003 .029 .052
Table IV. Multivariate analysis of postoperative AF Variables
Atrial index. The mean atrial indices for the patients with AF and without AF are shown. Pre, Prebypass period; post, postbypass period.
Age LA dimension BSA
OR (95% CI)
P value
1.11 (1.04-1.19) 1.74 (1.03-2.96) 114 (4.65-2810)
.002 .038 .004
Comparisons with previous studies ECG findings There were no significant differences in P-wave duration (112.6 ± 18.1 vs 114.5 ± 18.1 milliseconds, P = .89) or P-wave amplitude (1.09 ± 0.44 vs 1.09 ± 0.52 mV, P = .47) between the 2 groups.
Atrial index versus other clinical predictors By univariate analysis, increased age, lower atrial index, increased left atrial size, increased body surface area, longer aortic cross-clamp time, and postoperative myocardial infarction increased the odds of postoperative AF (Table III). All variables that were significant after univariate analysis were entered in a stepwise multivariate logistic regression model. In the final model, only age (OR 1.11, 95% CI 1.04–1.19, P = .002) and atrial dimension (OR 1.74, 95% CI 1.03–2.96, P = .038) were significant independent predictors of postoperative AF (Table IV). Body surface area also increased the odds of postoperative AF, but the CI was wide (OR 114, 95% CI 4.65–2810, P = .004), suggesting that larger sample size is necessary to confirm the accuracy of the estimate.
Discussion The major observations in our study are that advanced age and left atrial enlargement were independent predictors of AF after CABG. Although left atrial dysfunction was found to be a significant predictor of postoperative AF on univariate analysis, it did not remain as an independent predictor on multivariate analysis. P-wave analysis with standard 12-lead ECG did not provide predictive information of AF after CABG.
Our finding that 30.1% of patients had postoperative AF is similar to that reported by previous studies.1-3 Our results also agree with those from previous studies demonstrating that advanced age and LA enlargement were independent predictors of postoperative AF. In contrast, our study, along with others,6,16 did not confirm the finding that men have a higher risk of postoperative AF than women do.8,17 Whether sex is a surrogate of other predictors of postoperative AF remains to be determined.
LA dimension and function Although several studies had previously demonstrated that LA enlargement is predictive of AF after cardiac surgery, no previous study has investigated the role of atrial function in postoperative AF. Our study is the first to prospectively evaluate LA function by using intraoperative TEE as a predictor of AF after CABG. Our study was also the first to validate the utility of the LA index as a measurement of LA function. Our results demonstrated that the atrial index was significantly lower in the group who subsequently had postoperative AF than in those without postoperative AF, even before surgical intervention began, suggesting that pre-existist LA dysfunction may be a precursor for postoperative AF, independent of left ventricular function. After the operation, although atrial function was lower in patients who subsequently had postoperative AF compared with those without such complications, this difference did not reach statistical significance. Through atrial index and LA dimension analysis, it appears that the state of dilation of the atrium may be more important than the overall atrial function as represented by the total reservoir function (LAEF) and stroke distance (velocity time integral of the E and A waves) in predicting postoperative AF.
American Heart Journal Volume 143, Number 1
P-wave duration and P-wave amplitude Other studies found that prolonged P-wave duration was an independent predictor of AF after CABG.17 However, our study did not show significant difference in P-wave characteristics between groups with and without postoperative AF. Whether more detailed analysis such as signal-averaged P-wave duration will be of more predictive value will need to be determined by future studies.
Mechanism of postoperative AF The mechanism of postoperative AF is unclear. Many etiologic causes have been proposed, including inflammatory response causing atrial edema, pericarditis, or reperfusion injury8; atrial hypothermia resulting in prolonged conduction time and inducibility of AF because of shorter effective refractory period (ERP), dispersion of ERP, and slower conduction18; and hypoxic, ischemic, or hypokalemic changes resulting in shortening of the monophasic action potential duration.19 In addition, many studies also reported that advanced age is strongly associated with postoperative AF, as also demonstrated in the general nonsurgical population with AF.20 Our current results along with previous observations suggest that pre-existing age-related atrial remodeling, as evidenced by LA dilatation and decreased LA function, superimposed by the acute stress of surgical intervention, may have resulted in atrial ischemia and acute atrial failure, manifested clinically as postoperative AF.
Clinical relevance The current study demonstrated that LA dimension appeared to be more important than LA function in predicting postoperative AF. If LA function was relatively preserved in those who subsequently had postoperative AF, therapeutic regimens aimed at optimizing loading conditions of the atrium to preserve function may be important. These results suggest that conventional drugs such as angiotensin-converting enzyme inhibitors, which have been found beneficial in modulating ventricular remodeling in heart failure as well as chronic AF,21 may have an important impact in preventing AF through modulation of the atrial loading conditions. The use of preoperative or intrapreoperative echocardiographic measurement of LA dimension and function may have incremental value in identifying high-risk subjects, thus increasing the cost-effectiveness of prophylactic antiarrhythmic therapy.
Study limitations Our study focused only on the in-hospital period. In this study, the onset of the first AF episodes occurred predominantly in the early postoperative period (24-72 hours postoperatively), although episodes of AF can occur late in the in-hospital postoperative period. As a
Nakai et al 185
result, our study may have underestimated the incidence of postoperative AF.
Conclusions Advanced age and left atrial enlargement are independent predictors of AF after CABG. Our study also suggests that LA dimension appears to be more important than atrial function in predicting AF after CABG. Further study with a larger sample size will be necessary to confirm the independent predictive value of LA function. We thank Darwin Pastor and Kawalpreet Manku, MBBS, for assistance in data collection and analysis. We also thank the physicians in the Department of Cardiovascular Anesthesia and Department of Cardiovascular Surgery, Kaiser Permanente Medical Center, San Francisco, for their cooperation in making the study possible.
References 1. Creswell LL, Sschuessler RB, Rosenbloom M, et al. Hazards of postoperative atrial arrhythmias. Ann Thorac Surg 1993;56:539-49. 2. Mathew JP, Parks R, Savino JS, et al. Atrial fibrillation following coronary artery bypass graft surgery: predictors, outcomes, and resource utilization. Multicenter Study of Perioperative Ischemia Research Group. JAMA 1996;276:300-6. 3. Aranki SF, Shaw DP, Adams DH, et al. Predictors of atrial fibrillation after coronary artery surgery. Circulation 1996;94:390-7. 4. Taylor GJ, Malik SA, Colliver JA, et al. Usefulness of atrial fibrillation as a predictor of stroke after isolated coronary bypass grafting. Am J Cardiol 1987;60:905-7. 5. Reed GL III, Singer DE, Picard EH, et al. Stroke following coronary artery bypass surgery: a case-control estimate of the risk from carotid bruits. N Engl J Med 1988;319:1246-50. 6. Borzak S, Tisdale JE, Amin NB, et al. Atrial fibrillation after bypass surgery: does the arrhythmia or the characteristics of the patients prolong hospital stay? Chest 1998;113:1489-91. 7. Buxton AE, Josephson ME. The role of p wave duration as a predictor of postoperative atrial arrhythmias. Chest 1981;80:68-73. 8. Fuller JA, Adams GG, Buxton B. Atrial fibrillation after coronary artery bypass grafting. J Thorac Cardiovasc Surg 1989;97:821-5. 9. Steinberg JS, Zelenkofske S, Wong SC, et al. Value of the p-wave signal-averaged ECG for predicting atrial fibrillation after cardiac surgery. Circulation 1993;88:2618-22. 10. Mendes LA, Connelly GP, Mckenney PA, et al. Right coronary artery stenosis: an independent predictor of atrial fibrillation after coronary artery bypass surgery. J Am Coll Cardiol 1995;25:198202. 11. Zaman AG, Archbold AA, Helft G, et al. Atrial fibrillation after coronary bypass surgery: a model for preoperative risk stratification. Circulation 2000;101:1403-8. 12. Kempf FC, Hedberg A, Molinoff P, et al. The effect of pharmacologic therapy on atrial beta-receptor density and postoperative atrial arrhythmias [abstract]. Circulation 1983;68(III Suppl):57. 13. Ali IM, Sanalla AA, Clark V. Beta-blocker effects on postoperative atrial fibrillation. Eur J Cardiothorac Surg 1997;11:1154-7. 14. Frost L, Mortensen PE, Tingleff J, et al. Dofetilide Post-CABG Study Group: efficacy and safety of dofetilide, a new class III antiarrhyth-
American Heart Journal January 2002
186 Nakai et al
15.
16.
17.
18.
mic agent, in acute termination of atrial fibrillation or flutter after coronary bypass surgery. Int J Cardiol 1997;58:135-40. Hourigan LA, Schiller NB. The composite index of left atrial function: a noninvasive tool for quantitating left atrial function [abstract]. J Am Coll Cardiol 2000;35(A Suppl):458A. Crosby LH, Pifalo WB, Woll KR, et al. Risk factors for atrial fibrillation after coronary artery bypass grafting. Am J Cardiol 1990;66: 1520-2. Chang CM, Lee SH, Lu MJ, et al. The role of P wave in prediction of atrial fibrillation after coronary artery surgery. Int J Cardiol 1999; 68:303-8. Sato S, Yamaguchi S, Schuessler RB, et al. The effect of augmented
atrial hypothermia on atrial refractory period, condition, and atrial flutter/fibrillation in the canine heart. J Thorac Cardiovasc Surg 1992;104:297-306. 19. Pichlmaier AM, Lang V, Harringer W, et al. Prediction of the onset of atrial fibrillation after cardiac surgery using the monophasic action potential. Heart 1998;80:467-72. 20. Furberg CD, Psaty BM, Manolio TA, et al. Prevalence of atrial fibrillation in elderly subjects (the Cardiovascular Health Study). Am J Cardiol 1994;74:236-41. 21. Van Den Berg MP, Crijns HJGM, Van Veldhuisen DJ. Effects of lisinopril in patients with heart failure and chronic atrial fibrillation. J Card Fail 1995;1:355-64.
The following article is an AHJ Online Exclusive. Full text of this article is available at no charge at our Web site: www.mosby.com/ahj. Valvular and Congenital Heart Disease
Systemic platelet effects of contrast media: Implications for cardiologic research and clinical practice M. C. D. Dalby, MD,a S. J. Davidson, MD,b J. F. Burman, MD,b J. Clague, MD,a U. Sigwart, MD,a and S. W. Davies, MDa London, United Kingdom
Background Angiographic contrast media cause platelet activation and decrease aggregability in vitro. We have previously shown in vitro a significant antiplatelet effect of contrast media at the concentrations obtained locally in the coronary artery during angioplasty. It is not known, however, whether a systemic effect is present. Method Thirty patients undergoing diagnostic coronary angiography were prospectively randomized to receive the nonionic medium iohexol, ionic low-molecular-weight medium ioxaglate, or ionic highmolecular-weight medium diatrizoate. Platelet aggregability was measured before and after the investigation with whole blood electrical impedance aggregometry (WBEA) with collagen agonist and the PFA-100 (Dade, Miami, FL) platelet function analyzer with combined shear, collagen, and adenosine diphosphate as agonists. Results With WBEA, with iohexol no difference in impedance change was seen: (medians and ranges) before, 9.8 Ω (4.8-19.2 Ω) versus after, 9.6 Ω (2-19.2 Ω) (P not significant [NS]). With
ioxaglate a significant fall was seen: before, 8.6 Ω (6.4-15.2 Ω) versus after, 6.6 Ω (0-12.4 Ω) (P = .004). With diatrizoate a significant and greater fall was seen: before, 10.8 Ω (6.4-17.6 Ω) versus after, 6.6 Ω (0-10.8 Ω) (P = .002). With PFA, no difference in closure time was seen with any medium: iohexol before, 99 seconds (79-142 seconds) versus after, 142 seconds (63-128 seconds) (P NS); ioxaglate before, 120 seconds (75-258 seconds) versus after, 95 seconds (74258 seconds) (P NS); and diatrizoate before, 114.5 seconds (65250 seconds) versus after, 100.5 seconds (72-300 seconds) (P NS).
Conclusions Ionic but not nonionic contrast media have a systemic antiplatelet effect at diagnostic angiographic doses when measured with WBEA. Such an effect has not been shown before. This may explain the observed improved clinical outcome with ionic contrast media but also might confound platelet studies in coronary angioplasty. (Am Heart J 2002;143:e1.) 4/90/119998