| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Ann Thorac Surg 2002;73:1380-1386
© 2002 The Society of Thoracic Surgeons
a Department of Thoracic and Cardiovascular Surgery, University Hospital, Uppsala, Sweden
b Department of Medical Epidemiology, Karolinska Institute, Stockholm, Sweden
c Unit of Clinical Epidemiology, Department of Medicine, Karolinska Hospital, Karolinska Institute, Stockholm, Sweden
d Division of Cardiovascular Epidemiology, Karolinska Institute, National Institute of Environmental Medicine, Stockholm, Sweden
e Cardiovascular Laboratory, Thoracic Department, Karolinska Hospital, Stockholm, Sweden
Accepted for publication January 21, 2002.
* Address reprint requests to Dr St
hle, Department of Thoracic and Cardiovascular Surgery, University Hospital, S-751 85 Uppsala, Sweden
e-mail: elisabeth.stahle{at}thorax.uas.lul.se
 |
Abstract |
|---|
 Top Abstract IntroductionMaterial and methodsResultsCommentAcknowledgmentsReferences |
|---|
Methods. Between 1987 and 1996, 7,493 patients were discharged alive after primary coronary artery bypass grafting and were followed up to the first readmission, date of death, or December 31, 1996. The hazard ratios for the risk factors found were used to calculate a readmission risk score.
Results. A total of 4,780 (63.8%) patients were readmitted. The fraction not readmitted within 1, 5, and 10 years were 61%, 29%, and 14% (95% confidence intervals = 60 to 62, 28 to 30, and 12 to 16), respectively. The risk of readmission was highest early after operation and then gradually decreased. Older age, active smoking, diabetes, previous myocardial infarction, unstable angina, dyspnea, severe left ventricular dysfunction, advanced New York Heart Association functional class, bypass time of 2 hours or more, and length of stay all independently increased the risk of readmission. Female sex was a significant risk factor in univariate but not in multivariate analysis. In all age groups, women had a one unit higher risk score. Given the same risk score, the risk of readmission was similar between sexes.
Conclusions. This study showed an acceptable risk of readmission after coronary artery bypass grafting. Women more often had risk factors related to readmission. However, given identical disease severity, the risk was similar in men and women.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentAcknowledgmentsReferences |
|---|
Randomized controlled trials to evaluate the benefit of CABG procedures [1] have provided important information on the recurrence of anginal symptoms, the occurrence of myocardial infarctions, and the quality of life after the procedures. However, most of these trials are old and the patients included were in many ways highly selected. The results of these earlier studies cannot be applied to an unselected patient population undergoing CABG today. Moreover, the randomized trials included few women. A number of reports suggest that the outcome of CABG is not as favorable in women as in men [2–4]. Such information might possibly lead to delayed referral of women for CABG [5].
The postoperative readmission rate is one potential indicator of the long-term outcome of CABG. Although cause-specific readmission rates may be a more precise measure than total readmissions, the use of such an end point could introduce possible misclassification bias. The assignment of a diagnosis to an episode of readmission is not totally objective. In such cases, the diagnosis is not only based on symptoms and objective findings but is just as much left to the judgment of the physician. In patients who have undergone CABG, the reason for readmission may be considered to be related to coronary artery disease. Analyses of all-cause readmissions provide a more reproducible assessment of the total postoperative morbidity.
The number of reports on readmission rates after CABG has been limited. Moreover, most of those studies focused on the rate of early readmission as a measure of the quality of medical care at the primary (index) hospitalization for CABG [6–9]. As yet there have been no reports on readmission rates in a long-term perspective. The aim of the present study was to analyze the long-term morbidity after CABG as assessed by all-cause hospital readmissions, and to elucidate the effect of gender on the outcome. We identified predictors of the first hospital readmission after primary CABG during a 10-year period in the catchment area of a large Swedish university hospital.
|
Material and methods |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentAcknowledgmentsReferences |
|---|
Data collection
Since 1970, clinical data on all patients undergoing open-heart procedures at the department have been collected prospectively. For each of the patient studied, preoperative information concerning the extent and severity of each patient’s coronary artery disease and perioperative variables was collected from this database [10].
A family history of ischemic heart disease was defined as occurrence of such disease in parents/siblings with onset before the age of 60 years; hypertension meant antihypertensive medication or blood pressure 
140/90 mm Hg when admitted; and diabetes mellitus implied treatment with insulin or drugs. Left ventricular function was categorized as: intact, ejection fraction (EF) 0.50; moderate dysfunction, EF 0.30 to 0.49; or severe dysfunction, EF less than 0.30.
Follow-up and outcome
A unique 10-digit national registration number is allocated to every Swedish citizen. In January 1997 all patients were evaluated with respect to survival by computerized linkage to two national registers, namely a continuously updated census register and the Swedish Cause of Death Register. By use of these combined registers, all patients could be identified as being alive on December 31, 1996, or assigned a date of death. Late mortality was defined as death from any cause.
Hospital readmissions for any cause, used as an indicator of morbidity after CABG, were collected by computerized linkage with the National In-Patient Register (Center of Epidemiology, National Board of Health and Welfare). This register contains data of all hospital admissions throughout all Swedish hospitals. All patients were identified at the time of the CABG procedure, that is, index admission, and followed up through December 31, 1996. The mean follow-up period was 2.1 years, corresponding to a total of 15,673 patient-years. All diagnoses were registered according to the ICD9 classification (Swedish version) during the study period.
Statistical methods
The Kaplan-Meier method was used to calculate life-table estimates for death and first readmission. Because 66 patients died outside hospital without previous readmission, an end point defined as readmission or death outside hospital was also considered. However, that analysis gave results similar to the presented hazard ratios.
Uni- and multivariate analyses of factors related to the risk of first readmission were based on an age-stratified Cox proportional hazards model. The time at risk of readmission after the CABG was accumulated from the date of discharge of the index episode to either the date of first readmission, date of death, or December 31, 1996. All variables were then entered into the initial multivariate model and a final model was selected by means of stepwise regression. A cut-off level of p = 0.10 was used as a criterion for both inclusion in and removal of variables from the model. Forward, backward, and stepwise elimination yielded the same final model. The hazard ratio is presented as a measure of the risk of readmisssion in different groups. Separate models were estimated for 1-year intervals after discharge.
The final multivariate model was also used to calculate a readmission risk score, where the weights for each variable included were taken to be proportional to the excess hazard ratios in the final multivariate model, rounded to the first decimal: 
.
The readmission risk score was used as a single continuous or categorized predictor of readmission. The categorization was applied to check the assumption of linearity. Furthermore, the covariates not included in the final model were included together with the score to check the model selection. We also sequentially added two-factor interaction terms between the risk score and the individual risk factors, to check for effect modification. These analyses yielded no modification of the proposed model. Finally, the time dependence of the risk score was assessed by separate analyses in 1-year time intervals after discharge.
The impact of the score for men and women was evaluated separately in a Cox regression analysis. The readmission risk score was also considered as a dependent variable, characterizing the patient at discharge. Differences in risk score according to age at operation, gender, and year of operation were evaluated by analysis of variance.
|
Results |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentAcknowledgmentsReferences |
|---|
|
|
|
|
Use of the readmission risk score as a single predictor of readmission yielded an excess relative hazard per unit score of 8.9% (95% CI = 8.0 to 9.9), and the hazard ratio for risk of readmission increased linearily with the risk score (Table 4).
|
Stratified by gender, the excess relative hazard per unit readmission risk score was 9.3% (95% CI = 7.1 to 11.4) for women and 8.7% (95% CI = 7.6 to 9.9) for men, a difference that was nonsignificant. The risk of readmission, given the same score, was thus essentially the same for men and women.
On an average, women had a one unit (95% CI = 0.8 to 1.2) higher readmission risk score at discharge than men (Table 5). This difference was consistent over all age groups and did not change significantly with calendar year. Both age at operation and gender significantly influenced the mean readmission risk score (Table 5) (p < 0.0001), but no significant interaction between these factors was seen (p = 0.22).
|
Unstable angina, atrial fibrillation, dyspnea, a long bypass time, a high NYHA class, and severely impaired ventricular function all had a somewhat higher impact on the readmission rate within the first year than after 1 year, whereas the presence of diabetes mellitus showed increased influence with time after discharge.
|
Comment |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentAcknowledgmentsReferences |
|---|
Operations, CABG included, are well-known risk factors per se for later readmissions. In patients undergoing CABG, the coronary artery disease is already established and will develop further. In these patients, residual angina, myocardial infarctions, arrhythmias, and congestive heart failure are likely to lead to hospitalizations. We found that cardiac causes and precordial pain were the most common reasons and together accounted for about 35% of the first hospital readmissions on all occasions, irrespective of the length of time after discharge.
Unfortunately, this study does not provide information on the possible impact of secondary preventive treatments. More effective secondary prevention, including aggressive lipid-lowering treatment, might have affected the results further. Statins were introduced during the 1990s, but it was not until the Scandinavian Simvastatin Survival Study 45 study [11] was published in 1994 that these drugs were commonly used in patients with angina or previous myocardial infarction. Since then the target level of serum cholesterol has gradually been lowered in patients with manifest coronary artery disease. The implementation of the use of statins in post-CABG care has been rather slow in Sweden. Although most patients referred for CABG today are treated with statins, it may be assumed that the majority of the patients in the current study had not received an optimal lipid-lowering drug regimen.
An important issue is whether all-cause readmission rates provide valid information about the morbidity and quality of life after CABG. In this study we used the patient’s first readmission from any cause as an end point. Although cause-specific readmissions could provide additional information, their use might also introduce a number of potential biases. All-cause readmission, on the other hand, is easily defined and the necessary data are readily available. The indications for readmissions may have been altered during the study period. In addition, the type of health care system offering the operation, and the organization of the postoperative care and rehabilitation, may influence the probability of later rehospitalizations.
Few studies have dealt with readmission rates in a long-term perspective, but a number of analyses have provided information about short- and intermediate-term rates. Rates of readmission within 3 months of about 20% have been reported [7, 9] and within 2 years of approximately 40%. These figures are comparable to the rates in the present study. In those studies, emergency operation was an independent risk factor. However, estimates of left ventricular function also contributed [7, 9]. The risk factors related to risk of first readmission in the present study correspond to the factors known to influence mortality and also morbidity after CABG [12, 13]. Young patients (ie, younger than 50 years) had higher risks than those of older age groups, but were at the same risk level as the oldest age group. Left heart failure, which is a known risk factor for early morbidity and mortality after CABG [1, 10, 12, 13], was also an important risk factor for readmission. The use of the internal mammary artery reduced the risk of readmission in the univariate analysis and tended to do so also in the multivariate analysis. It can be assumed that the independent effect of the internal mammary artery will become more evident with prolonged follow-up [14].
Numerous studies have addressed the question of an interactive effect between gender and outcome of CABG, most of them indicating inferior results in women [2–4]. Early mortality has consistently tended to be higher among women. However, the long-term benefit has been shown to be similar among women and men in some studies [1, 15–17]. In the present study, female gender was a significant risk factor for readmission, with a 16% increase in risk in the univariate analysis. However, after correction for the effect of other risk factors in the multivariate analysis, the effect of female gender was reduced, and no longer statistically significant. It has been suggested that compared to men, women who undergo CABG have more comorbidities and tend to be older at the time of operation [5, 18]. This was confirmed in the present study, in which women consistently had higher risk scores at discharge. However, given the same preoperative disease severity, women and men have equal outcomes. Our finding that within each age group, women had an average of a one unit higher risk score may indicate that women are referred for CABG later and with more severe coronary artery disease than men.
In conclusion, this study showed an acceptable risk of readmission after CABG. Women more frequently had risk factors related to readmission. However, given identical disease severity, the risk was similar in men and women.
|
Acknowledgments |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentAcknowledgmentsReferences |
|---|
|
References |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentAcknowledgmentsReferences |
|---|
hle E., Bergström R., Edlund B., et al. Influence of left ventricular function on survival after coronary artery bypass grafting. Ann Thorac Surg 1997;64:437-444.
., Karlsson T., Herlitz J. Female sex is associated with increased mortality and morbidity early, but not late, after coronary artery bypass grafting. Eur Heart J 1996;17:1426-1431.This article has been cited by other articles:
|
|
 |
|
  S. J. Durham and J. P. Gold Late Complications of Cardiac Surgery Card. Surg. Adult, January 1, 2008; 3(2008): 535 - 548. [Full Text]
|
|
|
|
 |
|
 V. Guru, S. E. Fremes, P. C. Austin, E. H. Blackstone, and J. V. Tu Gender Differences in Outcomes After Hospital Discharge From Coronary Artery Bypass Grafting Circulation, January 31, 2006; 113(4): 507 - 516. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. L. Hannan, C. Wu, G. Walford, S. B. King III, D. R. Holmes Jr, J. A. Ambrose, S. Sharma, S. Katz, L. T. Clark, and R. H. Jones Volume-Outcome Relationships for Percutaneous Coronary Interventions in the Stent Era Circulation, August 23, 2005; 112(8): 1171 - 1179. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. A. Fox and N. A. Nussmeier Does Gender Influence the Likelihood or Types of Complications Following Cardiac Surgery? Seminars in Cardiothoracic and Vascular Anesthesia, December 1, 2004; 8(4): 283 - 295. [Abstract] [PDF]
|
|
|
|
|
|
V. Vaccarino, Z. Q. Lin, S. V. Kasl, J. A. Mattera, S. A. Roumanis, J. L. Abramson, and H. M. Krumholz Sex Differences in Health Status After Coronary Artery Bypass Surgery Circulation, November 25, 2003; 108(21): 2642 - 2647. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. L. Hannan, M. J. Racz, G. Walford, T. J. Ryan, O. W. Isom, E. Bennett, and R. H. Jones Predictors of Readmission for Complications of Coronary Artery Bypass Graft Surgery JAMA, August 13, 2003; 290(6): 773 - 780. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |
