|
|
||||||||
a Maine Medical Center, Portland, Maine
b The Dartmouth Institute for Health Policy and Clinical Practice, Section of Cardiology, Dartmouth Medical School and Dartmouth-Hitchcock Medical Center, Lebanon, New Hampshire
c The Dartmouth Institute for Health Policy and Clinical Practice, Lebanon, New Hampshire
Accepted for publication April 7, 2010.
* Address correspondence to Dr Kramer, Division of Cardiothoracic Surgery, Maine Medical Center R8, 22 Bramhall St, Portland, ME 04102 (Email: kramer{at}mmc.org).
Presented at the Forty-sixth Annual Meeting of The Society of Thoracic Surgeons, Fort Lauderdale, FL, Jan 25–27, 2010.
| Abstract |
|---|
|
|
|---|
Methods: We prospectively enrolled 668 nonemergent adult cardiac surgical cases. Patients having heart catheterization were divided into two groups: cardiac catheterization followed by cardiac surgery during the same hospital admission (same admission) or catheterization followed by surgery during a later admission (later admission). The AKI was defined by an increase in serum creatinine from baseline by 50% or greater or 0.3 (mg/dL) or greater. Univariable and multivariable logistic regression and propensity-matched analyses were conducted.
Results: The incidence of AKI was significantly higher in the patients who had same admission cardiac catheterization and surgery (50.2%) compared with patients who had surgery on a later admission (33.7%, p = 0.009). The adjusted odds ratio for surgery on a later admission was 1.54 (95% confidence interval: 1.11 to 2.13) suggesting a 54% increased risk of AKI. Propensity-matched results were similar with 1.58 (95% confidence interval: 1.13 to 2.22).
Conclusions: When cardiac catheterization and cardiac surgery occur during the same hospitalization, there is an increased risk for postoperative AKI. After cardiac catheterization, discharge and readmission for nonurgent surgery should be considered as such an approach might reduce the risk of AKI.
| Introduction |
|---|
|
|
|---|
| CARDIOTHORACIC ANESTHESIOLOGY:
The Annals of Thoracic Surgery CME Program is located online at http://cme.ctsnetjournals.org. To take the CME activity related to this article, you must have either an STS member or an individual non-member subscription to the journal.
|
It is common practice to provide same admission diagnostic and surgical services for patients undergoing cardiac surgery. In the absence of compelling reasons for surgery such as instability or life-threatening anatomy, discharge and subsequent readmission may be indicated. However, patients often undergo surgery after cardiac catheterization during the same admission for nonurgent reasons. The practice of operating on patients with a same-admission cardiac catheterization may induce a higher risk for acute kidney injury (AKI) and therefore an increased likelihood of morbidity and mortality [1–8].
Cardiopulmonary bypass and other intraoperative practices may contribute or be responsible for AKI through a variety of mechanisms such as hemodynamic alterations, ischemic injury secondary to generation of macroscopic and microscopic emboli, exposure of patient's blood components to the bypass circuit inducing a systemic inflammatory response syndrome, activation of neutrophils and free hemoglobin released from injured erythrocytes causing oxidative injury, and ischemia-reperfusion injury [9].
In addition to the intraoperative events noted above, there are preoperative and postoperative events that contribute to AKI in the cardiac surgical patient. A potentially modifiable risk for AKI could be the timing of cardiac surgery after cardiac catheterization. Acute kidney injury due to radiocontrast exposure is a well known complication of cardiac catheterization and is associated with increased mortality [10]. Heart catheterization and heart surgery in close proximity may pose additive risks for AKI when performed during the same admission. Ranucci and colleagues [7] and Medalion and colleagues [5] have shown the influence of timing of cardiac catheterization on the incidence of acute renal failure after cardiac surgery. Medalion and colleagues concluded that coronary artery bypass grafting should be delayed for at least 5 days in patients who receive a high contrast dose. Ranucci and colleagues concluded that delaying cardiac surgery beyond 24 hours of exposure to contrast has the potential to decrease the incidence of acute renal failure in elective cardiac surgery. However, it is not known whether cardiac surgery conducted on a later admission, compared with surgery during the same admission as for cardiac catheterization, is associated with a lower incidence of postoperative AKI.
Therefore, we prospectively collected catheterization and surgery data during 2008 at a single center. We analyzed the timing of cardiac catheterization and cardiac surgery to evaluate if same admission surgery is associated with an increased risk for AKI when compared with later admission surgery.
| Material and Methods |
|---|
|
|
|---|
We prospectively enrolled 722 consecutive adult cardiac surgical cases at a single NNECDSG center between January 2008 and December 2008. Patients without a preoperative catheterization (n = 8), with a history of dialysis treatment prior to surgery (n = 5), or undergoing emergent cardiac surgery (n = 41) were excluded from the analysis. The final cohort included 668 nonemergent patients. Patients undergoing cardiac catheterization were divided into two groups: catheterization followed by surgery during the same hospital admission (same admission) or on a later admission (later admission). Patients undergoing catheterization as an in-patient at an outside hospital and transferred during that admission to the surgical facility were considered same admission surgery. Patients undergoing catheterization at the surgical facility on a previous admission or if the catheterization was conducted at an outside hospital on a previous admission (not transferred) were considered a previous admission catheterization.
The most recent preoperative serum creatinine was measured at baseline prior to cardiac surgery, and 100% capture of daily postoperative serum creatinine up to 48 hours, then per clinician judgment until discharge. Acute kidney injury was defined prior to discharge from cardiac surgery by a 50% or greater or 0.3 or greater (mg/dL) increase in serum creatinine from baseline during the postsurgical hospitalization [11].
Statistical Analysis
Univariable and multivariable logistic regression was used to evaluate the relationship between same admission catheterization and previous admission catheterization. A nonparsimonious logistic regression model was used to identify significant baseline covariates to be used in the final adjustment using a backwards step approach and a p value 0.10 or less. The final multivariate logistic regression model adjusted for age, presence of coronary artery disease, baseline serum creatinine (mg/dL), pump time (minutes), and preoperative hematocrit. All other variables were not significant (p value > 0.1) and were not included in the final model, including timing between cardiac catheterization and surgery (in days), preoperative intraaortic balloon pump, surgery type, prior surgery, prior percutaneous coronary intervention, and ejection fraction less than 0.40. A propensity-matched analysis was conducted using a predictive model for the exposure (same admission cardiac catheterization), including gender, body mass index, ejection fraction less than 0.40, preoperative white blood cell count greater than 12,000, urgent priority, and type of surgery resulting in 550 matched patients. Conditional logistic regression was used to conduct the propensity-matched analysis. Multivariable modeling was conducted on the full cohort (n = 668) with 100% complete data, with the following exceptions: 41 patients who did undergo cardiopulmonary bypass were missing recorded pump times (minutes). To correct for this missing data, we set the pump time to the mean time. Nine patients had missing ejection fraction and for the indicator variable ejection fraction less than 0.40 were set to the null. A priori subgroup analyses were conducted on important subgroups of patients. All analyses were conducted using Stata 11.0 (College Station, TX).
| Results |
|---|
|
|
|---|
|
|
|
| Comment |
|---|
|
|
|---|
Several studies [4, 8, 12–14] have demonstrated an association between AKI after cardiac surgery and increased morbidity, short-term and long-term mortality, intensive care unit and hospital lengths of stay, and utilization of resources. Other studies [4, 13] have also demonstrated that postoperative increases in serum creatinine by 25% or greater or 0.5 or greater (mg/dL) are predictive of short-term and long-term mortality.
Our study is different than some others in that it compares the association between same admission cardiac catheterization and cardiac surgery with separate admissions [5, 7]. The implication is that it is safe, and possibly beneficial in terms of renal protection, to send patients home after cardiac catheterization with plans for surgery during a subsequent admission. Others have demonstrated the harmful renal effects from radiocontrast dye, maligning the kidney through acute tubular necrosis and renal tubular cell apoptosis [15]. It is likely through this mechanism and necessity of time for renal cell recovery prior to inducing new insults from surgery that we begin to understand the association between a same admission catheterization and postsurgical AKI. A second explanation could be found in the aggressive medical management of the patient during single hospital stay between catheterization and surgery may subject the patient to pharmacologic overload of the kidneys or other hospital-acquired adverse events. In addition to the intraoperative events, there are preoperative and postoperative events that play a role in contributing to AKI in the cardiac surgical patient (Table 4).
|
There are limitations of this analysis. First, it was generated from a single-center database and while collected prospectively, was evaluated retrospectively. The prospective data collection was conducted using a veteran quality improvement registry, which undergoes validation. However, the analysis was conducted retrospectively and is limited to unmeasured confounding. Second, despite adjusting for significant baseline covariates, there may be other confounding covariates that would mark the urgent group for both increased mortality and incidence of AKI. In spite of these limitations, this analysis moves us in the direction of separating the catheterization and surgical admission in time when safe, thereby mitigating the potential for AKI and increased short-term and long-term mortality [1, 16, 17].
There are several strengths to this study. The cohort was created by consecutive enrollment of over 700 patients with a broad mix of comorbidities and gender, which permits adequate generalizability of our findings. The serum creatinines were collected prospectively and were determined frequently on all subjects as a standard of care. Consequently, there is no ascertainment bias for AKI in patients who were sicker or hospitalized for a longer duration. Finally, this is a modern prospective cohort where many evidence-based strategies such as blood and glucose management are the standard of care, adding further to the generalizability of the data when compared with other centers using evidence-based protocols [18–20].
There is no question that the urgent group is going to have higher risk patients and comorbidities that would more likely be risk factors for AKI as demonstrated in our analyses. Embedded in the urgent group may be patients who are subjectively urgent and could safely be managed during a readmission. Furthermore, there may also be patients in the mix who are there for scheduling convenience. Those patients who cannot be managed as outpatients and be readmitted would be well served to wait a few days in the hospital before surgery.
It appears that there is an increased risk of renal injury when cardiac catheterization and cardiac surgery occur during the same hospitalization. Clearly, there are many patients admitted for unstable, life-threatening or disabling conditions that require surgical management during the same admission as their cardiac catheterization. The risk of AKI in that group of patients is a component of the overall risk in the urgent patient population, which is defined as patients that cannot leave the hospital prior to surgery.
The NNECDSG registry has shown us that most of the mortalities in the isolated coronary artery bypass grafting cohort occur in the urgent population. This information may present the opportunity to operate on patients in an elective time frame. Some of the patients who have surgery and cardiac catheterization during the same hospitalization and who may be characterized as having a reason to be operated upon urgently, may truly belong in the elective (nonurgent) group. For example, some of the patients, especially those who are far from home, appreciate the convenience of having their diagnostic catheterization and surgery during the same hospital admission. When they are grouped with patients who are truly urgent, it appears that they may be at increased risk for AKI, with its attendant increase in long- term and short-term mortality.
In conclusion, this prospective observational analysis presents the opportunity to compare future urgent and nonurgent patients using a protocol that would clearly state indications for cardiac surgery during the same admission as cardiac catheterization; for example, such as acute coronary syndrome or some other unstable pattern, critical left main trunk stenosis, life-threatening anatomy, or critical aortic stenosis. Patients admitted for cardiac catheterization and cardiac surgery during the same admission for purposes of convenience, or less compelling reasons, could be discharged and readmitted at a future date.
Our findings suggest there is an opportunity to decrease the incidence of AKI, along with its long-term mortality, by separating the truly urgent patients from elective or nonurgent patients. The a priori subgroup analysis (Table 3) generates the hypothesis that conducting cardiac catheterization during a previous admission is renoprotective among elective patients, patients without a recent myocardial infarction (within 7 days of surgery), and among patients without unstable angina regardless of the type of surgery, congestive heart failure, or left main disease. Increasing the time interval between the heart catheterization by sending patients home after a cardiac catheterization may allow resolution of some of the predisposing factors contributing to AKI in the postoperative cardiac surgical patient and reduce overall morbidity and mortality for cardiac surgery patients.
| Discussion |
|---|
|
|
|---|
Dr Kramer, you and your colleagues have brought forth a complex and often vexing issue faced by your brethren in this society on a daily basis: when is the best time to operate on our patients? There are many medical and nonmedical issues that go into the decision-making process. We must balance the medical issues with competing interests, which often include the convenience of the patients and their families, referring physician pressures, administrative pressures to keep our operating rooms full and the balance sheet in the black, and our own pressures to keep our lifestyles acceptable and make efficient use of our time. Often, doing the right thing is neither easy nor apparent.
Using the resources of the Northern New England Cardiovascular Disease Study Group, you have prospectively collected data to assess the impact of performing cardiac surgery during the same admission as the patient's cardiac catheterization. Knowing the well-established relationship between acute kidney injury on morbidity, mortality, and the cost of cardiac surgery, you have looked at the impact of same admission surgery and acute kidney injury. You found that 45% of the 722 patients operated on during the year 2008 developed acute kidney injury with 50% of the patients developing acute kidney injury who had same admission surgery and only 33% of the patients in the delayed surgery group developing acute kidney injury. While further delineation between the medically and sociologically driven patients in the same admission group would have been interesting, your logistic regression demonstrates that acute kidney injury is equally likely to develop in patients with valvular, coronary artery, and combined diseases. With that in mind, there certainly have to be differential costs involved in discharging patients versus same hospitalization surgery. Have you had the opportunity to look at these differential costs and what are they? Is it more or less expensive to have surgery during the same hospitalization? And as a corollary of this economic question, are you able to tell us the specific reasons for same admission surgery in your cohort of patients who are nonmedically urgent and were these economic reasons enough to justify the additional risks to these patients?
Secondly, what is the optimal time to wait between catheterization and surgery? And as a corollary to this, was the quantity and type of contrast different in any of these groups and what message do we send our surgical and cardiological colleagues as to when is the best time to operate on patients?
Thank you, congratulations, and I look forward to your answers.
DR KRAMER: Thank you, Dr Conte. I am going to answer the second question first. The optimal time between cardiac catheterization and surgery was not revealed in our data. Ranucci and Medalion have both published papers regarding patients who had high volumes of contrast media during their heart catheterization. Their data suggested waiting between one and five days.
Since the acute kidney injury could be an epiphenomenon there may be resolution over time of other organ injuries we have yet to identify. In addition, anemia, drugs, inflammation, etc. could be addressed, increasing the benefit from waiting. We are going to advocate for a separate admission, not only to mitigate the potential for acute kidney injury but to potentially address other issues. I would suggest two weeks in the nonurgent patient and 5 days if the patient cannot leave the hospital.
With regard to your first question, we did not do a cost analysis, but in the value analysis, improved outcomes often are a way to decrease cost. There may be as much as a 15% survival disadvantage over five years when the patients have acute kidney injury, similar to the survival curves that have been published in the observational work on blood transfusions.
If we can move some of those patients into a later time frame where they were not exposed to that opportunity to have a tubular injury, they have an opportunity for better outcomes.
DR OMAR LATTOUF (Atlanta, GA): You have used a rise in creatinine of 0.3 as an indicator or a surrogate of acute kidney injury. We know that a rise in creatinine could be multifactorial in nature. Certainly intravascular contraction rather than renal injury could be a cause. How can you differentiate based on a small rise in creatinine that it is not intravascular contraction? Thank you.
DR KRAMER: That is a very good question. As you know, there has been a lot of literature addressing the definition of acute kidney injury. The AKIN (Acute Kidney Injury Network) criterion uses the 0.3 mg/dL or 50% increase in baseline creatinine in order to determine whether a patient had acute kidney injury. The difference between volume responsive acute kidney injury and renal tubular cell death is likely reflected in the duration of the creatinine elevation. A patient whose creatinine increases 0.3 mg/dL or 50% and comes down to baseline in a day may be in the category of patients who are volume responsive. In an analysis relating duration of creatinine elevation to long-term survival, patients whose creatinine was elevated for two days or more were less likely to survive five years.
In the group of patients who have a 0.3 mg/dL creatinine elevation over baseline for less than two days, there are probably some volume-responsive patients who did not have tubular cell death.
| Acknowledgments |
|---|
|
|
|---|
| References |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
J. R. Brown Invited Commentary Ann. Thorac. Surg., February 1, 2013; 95(2): 519 - 519. [Full Text] [PDF] |
||||
![]() |
R. S. Kramer, C. R. Herron, and J. Brown Association of catheterization with acute kidney injury in the cardiac surgical patient J. Thorac. Cardiovasc. Surg., January 1, 2013; 145(1): 312 - 313. [Full Text] [PDF] |
||||
![]() |
N. D. Andersen and G. C. Hughes Reply to the Editor J. Thorac. Cardiovasc. Surg., January 1, 2013; 145(1): 313 - 314. [Full Text] [PDF] |
||||
![]() |
N. D. Andersen, J. B. Williams, E. L. Fosbol, A. A. Shah, S. D. Bhattacharya, R. H. Mehta, and G. C. Hughes Cardiac catheterization within 1 to 3 days of proximal aortic surgery is not associated with increased postoperative acute kidney injury J. Thorac. Cardiovasc. Surg., June 1, 2012; 143(6): 1404 - 1410. [Abstract] [Full Text] [PDF] |
||||
![]() |
J. Hajj-Chahine, C. Jayle, J. Tomasi, H. Houmaida, and P. Corbi Same Admission Cardiac Catheterization-Cardiac Surgery: Increased Incidence of Acute Kidney Injury and Mediastinitis Ann. Thorac. Surg., August 1, 2011; 92(2): 776 - 776. [Full Text] [PDF] |
||||
![]() |
R. S. Kramer, R. D. Quinn, R. C. Groom, M. A. Kellett, and J. R. Brown Reply Ann. Thorac. Surg., August 1, 2011; 92(2): 776 - 777. [Full Text] [PDF] |
||||
![]() |
R. S. Kramer Invited Commentary Ann. Thorac. Surg., June 1, 2011; 91(6): 1796 - 1797. [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 |