|
| HOME | SUBSCRIPTIONS | CURRENT ISSUE | PAST ISSUES | CARDIOSOURCE | SEARCH | HELP | FEEDBACK |
|
|
|
|||||||||
|
|||||||||||
|
J Am Coll Cardiol, 2003; 42:1767-1776, doi:10.1016/j.jacc.2003.05.008 © 2003 by the American College of Cardiology Foundation |
* Department of Anesthesia, Basel, Switzerland
Department of Anesthesia Internal Medicine, Division of Cardiology, Basel, Switzerland
Department of Anesthesia Surgery, University of Basel/Kantonsspital, Basel, Switzerland
Manuscript received April 2, 2003; accepted May 7, 2003.
* Reprint requests and correspondence: Dr. Miodrag Filipovic, Department of Anesthesia, University of Basel/Kantonsspital, CH-4031 Basel, Switzerland.
mfilipovic{at}uhbs.ch
 |
Abstract |
|---|
 Top Abstract MethodsResultsDiscussionReferences |
|---|
BACKGROUND: In cardiac-risk patients undergoing major noncardiac surgery, the short- and long-term prognoses are mainly influenced by perioperative cardiac complications. Heart rate variability and cTnI are important prognostic markers in patients with congestive heart failure and myocardial infarction.
METHODS: In a prospective study, 173 patients with CAD or at high risk of CAD undergoing major noncardiac surgery were followed up for one year. The main outcome measure was all-cause mortality. In addition to clinical parameters and established risk scores, HRV and cTnI were assessed perioperatively.
RESULTS: Twenty-eight (16%) patients died within one year. Multivariate logistic regression analysis revealed three findings that were independently associated with death within the first year after surgery: the revised cardiac risk index (odds ratio 6.2 [95% confidence interval 1.6 to 25], depressed HRV before induction of anesthesia (16.2 [2.8 to 94]), and elevation of cTnI on postoperative day 1 or 2 (9.8 [3.0 to 32]).
CONCLUSION: Depressed HRV before induction of anesthesia and elevated cTnI postoperatively are independent and powerful predictors of one-year mortality for patients at risk of CAD undergoing major noncardiac surgery and add incremental prognostic information to established risk scores that only consider preoperative information.
| ||||||||||||||||||||||
Perioperative myocardial ischemia is one event strongly correlated with adverse long-term outcome (3,9,10). Perioperative myocardial ischemia occurs in up to 40% of patients at risk of CAD (10) but is usually clinically silent (11,12) and, accordingly, difficult to detect. Myocardial ischemia leading to myocardial cell death can be detected by specific markers of myocardial cell injury, for example, cardiac troponin I (cTnI) or T. Their prognostic values have been shown in nonsurgical patients with acute coronary syndromes (13–15). In surgical patients, cardiac troponin elevation was shown to detect perioperative myocardial infarction (16), to predict cardiac complications (17), and to predict six-month outcome (4,18). Depressed heart rate variability (HRV) as a parameter of the patient's autonomic function was shown to be another predictor of future cardiac events, cardiac death, and all-cause mortality in patients with acute myocardial infarction and unstable angina (19–23). Reduced HRV was also strongly associated with congestive heart failure (24–26). However, its prognostic value in surgical patients was evaluated only in a small number of patients for short-term outcome (27).
The aim of our study was to evaluate whether reduced HRV and perioperatively elevated levels of cTnI are predictive for long-term mortality and provide additional information to established risk scores in patients with documented CAD or at high risk of CAD undergoing major noncardiac surgery.
|
Methods |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
|
Renal failure was defined as serum creatinine >177 µmol/l (= 2 mg/dl) (8). Diabetes mellitus was diagnosed if the patient was treated with insulin and/or oral antidiabetic agents. History of congestive heart failure was defined according to Lee et al. (8). Arterial hypertension was diagnosed in patients with concurrent antihypertensive therapy and in patients with systolic arterial pressure >160 mm Hg, and/or diastolic arterial pressure >95 mm Hg at admission.
Measurements. All data were collected prospectively. The entire perioperative management of the patients including preoperative evaluation, anesthetic technique, and cardiac and antihypertensive medication was guided by the physician in charge and not influenced by the study protocol. A study physician examined all patients before surgery and daily for the first six days after surgery.
Continuous electrocardiography (ECG) recordings (Marquette Series 8500, Marquette Electronics, Milwaukee, Wisconsin) were started before induction of anesthesia in the operating theatre and continued for the next 48 h. Three bipolar leads (V5, inverse Nehb J, aVF) were recorded (29). Tapes were processed on a computerized analyzer (Marquette Laser SXP). ST-segment trend analysis was manually reviewed on all three leads according to the published guidelines (30) by a physician blinded to all other study data.
For HRV measurements, the computer system classified the QRS complexes as normal beats, ectopic beats (supraventricular and ventricular), and artifacts. This classification was controlled visually on a screen and corrected manually when necessary. Special care was taken to eliminate from further analysis ectopic beats or artifacts that had been classified incorrectly as normal by the computer algorithm. Heart rate variability subsequently was calculated using commercially available software (Marquette, Version 5.8) according to published guidelines (31).
Time domain measurements of HRV (SD of normal-beat intervals, SD of the average of normal-beat intervals in 5-min segments of the entire recording, the square root of the mean of the sum of the squares of the differences between adjacent normal-beat intervals, and percent of normal-beat intervals differing by more than 50 ms in the entire recording) were calculated for the first and second 24-h period of recording according to current guidelines (31). Frequency domain measurements of HRV were calculated for both 24-h periods. In addition, short-term periods of 6 min were calculated at the following time points: baseline (immediately after start of the recording and before induction of anesthesia), and at 9 AM of the first and the second postoperative mornings. Using a fast Fourier transformation algorithm and a Hanning spectral window, the powers of the following frequency bands were calculated: total power (TP), (0.01 to 1.0 Hz); low frequency power (LF), (0.04 to 0.15); high frequency power (HF), (0.15 to 0.4 Hz); and the LF/HF ratio. Measurements of HRV were available in 148 of 173 (86%) patients and could not be obtained in 25 patients, 19 of whom had atrial fibrillation, three very frequent supraventricular premature beats, and one a pacemaker-triggered heart rhythm, whereas recordings could not be analyzed because of technical problems in two patients.
Cardiac troponin I was measured before surgery, immediately after arrival in the postanesthesia care unit, 8 h after the end of surgery, daily for the first three days, and on day 6 after surgery. For cTnI analysis, a commercially available kit (AxSYM, Abbott, Abbott Park, Illinois) was used with the normal upper limit of 2 µg/l specified by the manufacturer.
Twelve-lead ECGs were performed preoperatively, postoperatively after arrival in the postanesthesia care unit, and on postoperative days 1 and 6. They were analyzed according to the Minnesota criteria (32) by an investigator blinded to all other study data.
Follow-up and outcome measurements. Semistructured follow-up interviews were done by telephone 6 and 12 months postoperatively to determine mortality and establish reasons for death. All reported events were verified by the family physicians of the patients or by a review of hospital charts. However, the correct cause of death could not be unambiguously defined in some patients who died outside of a hospital.
Statistical analysis. The Mann-Whitney U test was used to compare HRV data. Differences in categorical variables were assessed by the Fisher exact test. Cutoff values for prediction of one-year all-cause mortality for cTnI and for the LF/HF ratio were based on receiver operating characteristics (ROC) (33). Potential univariate parameters of adverse outcome were identified using logistic regression. A first multivariate logistic regression analysis was performed by entering baseline variables with an initial p value of <0.10 into the model. A second multivariate logistic regression analysis was performed by entering into the model the two investigated perioperative findings (cTnI elevation, LF/HF depression) along with the baseline scoring that had the highest prognostic power (revised cardiac risk index). In addition, Kaplan-Meier survival curves were compared with the log-rank test. All analyses were performed using StatView computer package version 5.0 (SAS Institute Inc., Cary, North Carolina). A p value <0.05 was considered to indicate a significant difference.
|
Results |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
|
|
|
|
In addition to the above-mentioned findings, HRV and cTnI were strongly associated with death. In all analyzed long- and short-term periods, the LF/HF ratio of survivors was significantly higher compared with nonsurvivors. Seventy-nine patients (53%) had an LF/HF ratio <2 in the preoperative short-term measurement, 19 (24%) of whom died. Sixty-nine patients (47%) had an LF/HF ratio >2, two (3%) of whom died (Fig. 1B, for the corresponding Kaplan-Meier survival curve). The area under the ROC curve for LF/HF was 0.76. An LF/HF ratio <2 in the baseline measurement was strongly associated with death after one year (OR 10.6, 95% CI 2.4 to 48, p < 0.0001). The associations between one-year mortality and the LF/HF ratios on the first and second postoperative mornings were also significant but slightly weaker than that of the baseline measurement (OR 4.6, 95% CI 1.3 to 17, p = 0.02; OR 6.3, 95% CI 1.4 to 29, p = 0.02, respectively). In addition to the LF/HF ratio, the SD of normal-beat intervals, and the SD of the average of normal-beat intervals in 5-min segments of the entire HRV recording were lower in the first 24 h of recording in the patients who died compared with those who survived. In the short-term measurements, TP at baseline was lower in survivors compared with nonsurvivors. All other HRV measurements were statistically similar in these two groups. The LF/HF ratio at baseline was lower in patients with a history of congestive heart failure and in patients with renal failure. However, no differences were found between patients with or without diabetes, hypertension, or known or suspected CAD.
Cardiac troponin I was elevated (>2.0 µg/l) on the first and/or second postoperative mornings in 27 (16%) patients, 12 (44%) of whom died (Fig. 1C, for the corresponding Kaplan-Meier survival curve). The area under the ROC curve for cTnI elevation was 0.74. An elevation of cTnI >2 µg/l on the first and/or second postoperative mornings was strongly associated with 12-month all-cause mortality (OR 6.5, 95% CI 2.6 to 16, p < 0.0001). Cardiac troponin I was elevated preoperatively in 6 patients (4 died), immediately postoperatively in 6 patients (4 died), 8 h postoperatively in 9 patients (5 died), on the first postoperative morning in 11 patients (7 died), on the second in 24 patients (11 died), on the third in 22 patients (7 died), and on the sixth in 13 patients (3 died). Cardiac troponin I was elevated preoperatively, but not postoperatively, in only one patient who had a good outcome.
Of the 27 patients who had a cTnI elevation on the first and/or second postoperative mornings, 16 had an LF/HF ratio <2 at baseline, of whom 10 patients (63%) died. In contrast, only one patient (2%) died of the 62 patients who did not have a cTnI elevation and had an LF/HF ratio >2.
Additional findings associated with one-year mortality were regional anesthesia compared with general or combined regional/general anesthesia, evidence of myocardial ischemia in the Holter ECG recordings, and a rise in creatine kinase MB mass. In contrast, gender, histories of stroke, coronary revascularization, and myocardial infarction or evidence of perioperative myocardial infarction in the 12-lead ECG were not associated with death. Separate univariate logistic regression analysis results for patients with documented CAD and patients with high risk for CAD (defined according to Table 1) are summarized in Table 5.
|
|
|
|
|
Discussion |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
Clinical scores to predict risk. Established risk scores use the patient's history and the preoperative clinical status to predict short-term outcome or to identify patients with the need for more detailed cardiac testing (28). The revised cardiac risk index (8) is the latest adaptation of the well-established Goldman score (5). This revised cardiac risk index (8) was a strong independent predictor of adverse outcome in the multivariate analysis, and even was marginally statistically significant in the univariate analysis. The index was developed originally to predict short-term morbidity and mortality (8). Therefore, application of this risk score to predict long-term outcome may need some adaptations, mainly the definition of "high-risk surgery." This definition does not include infrainguinal vascular procedures. This exclusion seems questionable because all patients undergoing vascular procedures are at a substantial risk (34,35).
Our results confirm the prognostic value of comorbidities such as history of renal insufficiency, congestive heart failure, and diabetes for long-term prognosis. This is not surprising because these diseases are elements of the revised cardiac risk index.
HRV parameters to predict risk. Our study identifies an LF/HF ratio <2 before induction of anesthesia as a strong and independent predictor of all-cause mortality after one year. In surgical patients, the prognostic value of HRV measurements has been shown only for short-term outcome in a small number of patients (27), but it has not been evaluated for long-term outcome. In contrast, the prognostic value of HRV in nonsurgical patients with acute myocardial infarction and unstable angina has been established in many studies (19,31,36). These studies consistently found a reduction in one or several parameters of HRV in patients with CAD compared with normal subjects (20,36). In patients with acute myocardial infarction, reduced HRV (e.g., a reduced LF/HF ratio [19]) predicted a worse outcome (19,21,22,37). The reduction in HRV mainly is explained by an imbalance of the sympathetic and parasympathetic nervous system (22,31). Reduced HRV, evident as a reduction in LF and the LF/HF ratio, was also found in patients with heart failure (24–26), and is explained by a central regulatory impairment (24). The reduced LF/HF ratio before induction of anesthesia in many of our patients and its strong association with subsequent bad outcome may indicate the presence of (clinically silent) congestive heart failure. The lower LF/HF ratio in patients with a history of congestive heart failure may further support this explanation. Postoperative HRV measurements did not provide additional information to preoperative HRV measurements for risk assessment, although the LF/HF ratio was predictive of outcome at all analyzed time points. Because the study recordings started only shortly before induction of anesthesia, our data do not reveal the onset of HRV reduction.
Cardiac markers to predict risk. As a third parameter, a perioperative increase in cTnI was found to be highly predictive of one-year mortality. This is in agreement with the findings of Lopez-Jimenez et al. (4). They demonstrated, in a study of 772 patients, a correlation between a rise in cTnT and cardiac events within six months after major noncardiac surgery (adjusted OR 4.6) (4). A recent study by Kim et al. (18) reported the association of a postoperative rise in cTnI with six-month mortality after vascular surgical procedures. In agreement with results of our study, both former studies found a strong association between troponin elevation and six-month outcome; the association was even stronger in our more selected population. These studies are in accordance with results from studies in nonsurgical patients with acute coronary syndromes that also found a correlation between troponin elevation and mortality (14,15). Accordingly, one can hypothesize that a substantial percentage of cardiac risk patients suffer during the perioperative period from the equivalent of an acute coronary syndrome. This acute coronary syndrome often remains clinically undetected but provokes myocardial cell damage with a rise in troponin. As in the nonsurgical population, myocardial ischemia with cell damage is associated with a worse prognosis. A strong association of perioperartive ischemia and long-term mortality is well-established (3,4,10) and is also in agreement with our findings.
The troponin analyses before surgery and on days 1 and 2 after surgery had the highest prognostic value and should be considered as a component of perioperative risk assessment. Further troponin analyses on the day of surgery and analyses later than two days after surgery did not add incremental predictive information.
Study limitations. Our study has some important limitations. First, its sample size was relatively small, but large enough to identify new independent risk predictors. However, these findings have to be confirmed in a second prospectively studied large patient cohort. Second, our results cannot be applied to patients with a lower risk profile or to patients undergoing minor surgery, because we investigated patients at high risk of CAD. A further important limitation is the timing of Holter recordings that were started only 1 to 2 h before surgery. If a reduction in HRV is used to identify patients at risk before surgery and to serve as a basis for ameliorating their outcome, HRV measurements should be performed days before surgery. In addition, all HRV data in our study were calculated from Holter recordings, a time-consuming process and not suitable for daily routine. However, there are commercially available ECG devices that perform frequency-domain analyses of HRV on-line within 5 to 10 min. Whether such on-line measurements contain the same prognostic information as the Holter analysis in our study needs to be confirmed.
Conclusions. In conclusion, our study identified three perioperative parameters as independent predictors of all-cause mortality within one year after major noncardiac surgery in patients with documented CAD or at high risk of CAD. These parameters are the revised cardiac risk index (8), representing the patient's status and comorbidities; the LF/HF ratio, representing the autonomic function immediately preoperatively; and a rise in cTnI, representing perioperative myocardial cell damage. The clinical value of these findings needs to be assessed in future studies. These studies will have to define concepts that use HRV and/or troponin to identify surgical patients who may benefit from (postoperative) cardiac reevaluation and optimization of cardiac therapy.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
|
References |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  T. Laitio, J. Jalonen, T. Kuusela, and H. Scheinin The Role of Heart Rate Variability in Risk Stratification for Adverse Postoperative Cardiac Events Anesth. Analg., December 1, 2007; 105(6): 1548 - 1560. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 NACB WRITING GROUP MEMBERS, A. H.B. Wu, A. S. Jaffe, F. S. Apple, R. L. Jesse, G. L. Francis, D. A. Morrow, L. K. Newby, J. Ravkilde, W.H. W. Tang, et al. National Academy of Clinical Biochemistry Laboratory Medicine Practice Guidelines: Use of Cardiac Troponin and B-Type Natriuretic Peptide or N-Terminal proB-Type Natriuretic Peptide for Etiologies Other than Acute Coronary Syndromes and Heart Failure Clin. Chem., December 1, 2007; 53(12): 2086 - 2096. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. A. Fleisher, J. A. Beckman, K. A. Brown, H. Calkins, E. L. Chaikof, K. E. Fleischmann, W. K. Freeman, J. B. Froehlich, E. K. Kasper, J. R. Kersten, et al. ACC/AHA 2007 Guidelines on Perioperative Cardiovascular Evaluation and Care for Noncardiac Surgery: A Report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (Writing Committee to Revise the 2002 Guidelines on Perioperative Cardiovascular Evaluation for Noncardiac Surgery) Developed in Collaboration With the American Society of Echocardiography, American Society of Nuclear Cardiology, Heart Rhythm Society, Society of Cardiovascular Anesthesiologists, Society for Cardiovascular Angiography and Interventions, Society for Vascular Medicine and Biology, and Society for Vascular Surgery J. Am. Coll. Cardiol., October 23, 2007; 50(17): e159 - e242. [Full Text] [PDF]
|
|
|
|
 |
|
 L. A. Fleisher, J. A. Beckman, K. A. Brown, H. Calkins, E. L. Chaikof, K. E. Fleischmann, W. K. Freeman, J. B. Froehlich, E. K. Kasper, J. R. Kersten, et al. ACC/AHA 2007 Guidelines on Perioperative Cardiovascular Evaluation and Care for Noncardiac Surgery: A Report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (Writing Committee to Revise the 2002 Guidelines on Perioperative Cardiovascular Evaluation for Noncardiac Surgery) Circulation, October 23, 2007; 116(17): e418 - e500. [Full Text] [PDF]
|
|
|
|
 |
|
 G. M. Howard-Alpe, J. W. Sear, and P. Foex Methods of detecting atherosclerosis in non-cardiac surgical patients; the role of biochemical markers Br. J. Anaesth., December 1, 2006; 97(6): 758 - 769. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P.J. Devereaux, L. Goldman, S. Yusuf, K. Gilbert, K. Leslie, and G. H. Guyatt Surveillance and prevention of major perioperative ischemic cardiac events in patients undergoing noncardiac surgery: a review Can. Med. Assoc. J., September 27, 2005; 173(7): 779 - 788. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. Bonnet and E. Marret Influence of anaesthetic and analgesic techniques on outcome after surgery Br. J. Anaesth., July 1, 2005; 95(1): 52 - 58. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. Hanss, D. Block, P. H. Tonner, B. Bein, T. Laitio, T. Kuusela, H. Huikuri, J. Jalonen, E. Kentala, T. Makikallio, et al. Prediction of Postoperative Myocardial Ischemia by Fractal Heart Rate Dynamics * Response Anesth. Analg., January 1, 2005; 100(1): 289 - 290. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | SUBSCRIPTIONS | CURRENT ISSUE | PAST ISSUES | CARDIOSOURCE | SEARCH | HELP | FEEDBACK |
