| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Department of Cardiac Surgery, Morristown Memorial and Gagnon Heart Hospital, Atlantic Health, Morristown, New Jersey
b Department of Cardiac Research, Morristown Memorial and Gagnon Heart Hospital, Atlantic Health, Morristown, New Jersey
c Department of Psychiatry, Morristown Memorial and Gagnon Heart Hospital, Atlantic Health, Morristown, New Jersey
d Office of Grants and Research, Biostatistics, Atlantic Health, Morristown, New Jersey
e Neuropsychology, Morris Psychological Group, Parsippany, New Jersey
Accepted for publication August 29, 2008.
* Address correspondence to Dr Slater, Mid-Atlantic Surgical Associates, 95 Madison Avenue, Morristown, NJ 07962 (Email: james.slater{at}atlantichealth.org).
Presented at the Forty-third Annual Meeting of The Society of Thoracic Surgeons, San Diego, CA, Jan 29–31, 2007.
| This article has been selected for the open discussion forum on the CTSNet Web Site: http://www.ctsnet.org/sections/newsandviews/discussions/index.html
|
 |
Abstract |
|---|
 Top Abstract IntroductionMaterial and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
Methods: The rSO 2 was monitored intraoperatively in a cohort of primary CABG patients. Patients were prospectively randomized to a blinded control group or an unblinded intervention group. Cognitive function was assessed preoperatively, postoperatively, and at 3 months using a battery of standardized neurocognitive tests. Cognitive decline was defined as a decrease of one standard deviation or more in performance on at least one neurocognitive measure. The rSO 2 desaturation score was calculated by multiplying rSO 2 below 50% by time (seconds). Multivariate logistic regression models were used to assess cognitive decline and hospital stay. The change in cognitive performance was also assessed using a multivariate linear regression model.
Results: Patients with rSO 2 desaturation score greater than 3,000%-second had a significantly higher risk of early postoperative cognitive decline [p = 0.024]. Patients with rSO 2 desaturation score greater than 3,000%-second also had a near threefold increased risk of prolonged hospital stay (>6 days) [p = 0.007].
Conclusions: Intraoperative cerebral oxygen desaturation is significantly associated with an increased risk of cognitive decline and prolonged hospital stay after CABG.
Numerous studies have placed the incidence of early postoperative neurocognitive decline at greater than 50% in the coronary artery bypass population [1–4]. Intraoperative cerebral ischemia and cerebral oxygen desaturation have been proposed as possible mechanisms of postoperative cognitive dysfunction [5–7]. Cerebral oximetry monitoring is being used increasingly to monitor frontal lobe perfusion during cardiac and noncardiac surgery. Noninvasive cerebral oximetry uses near-infrared reflectance spectroscopy to measure frontal lobe regional cortical oxygen saturation.
An association between cerebral oxygen desaturation during cardiac surgery and postoperative cognitive dysfunction, prolonged intensive care unit (ICU), and hospital stays has been demonstrated [6, 8]. However, these studies were not prospectively randomized and lacked a control group. In a blinded, randomized trial, Murkin and colleagues [9] demonstrated that treatment of cerebral oxygen desaturation prevented prolonged desaturation and was associated with a shorter ICU length of stay and a significantly reduced incidence of major organ morbidity and mortality.
In an effort to determine the usefulness and efficacy of intraoperative cerebral oximetry monitoring we designed a prospective randomized trial in a coronary artery bypass grafting (CABG) population. Our objective was to examine and describe a relationship between intraoperative cerebral oxygen desaturation and adverse events such as neurocognitive decline and increased length of stay.
|
Material and Methods |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
Patients were randomized to one of two groups, intervention or control, prior to their arrival in the operating room. Randomization was based on a table of random numbers. The first number was blindly chosen, subsequent assignments were sequentially dictated by the table. Even numbered patients were assigned to the intervention group and odd numbers to the control group. The patient was blinded to their group assignment.
Neurocognitive testing was performed preoperatively, prior to discharge and at 3 months. The neuropsychologic portion of the study design followed the consensus statement on assessment of neurobehavioral outcomes after cardiac surgery [10]. The neurologic assessment included the presence or absence of visual disturbance, aphasia, paralysis or weakness and mental status evaluation. Cognitive function was assessed using a battery of standardized neuropsychologic tests to evaluate the domains of attention, memory, manual dexterity, and frontal eye field dysfunction. Frontal lobe oculomotor activity was measured by saccadic and anti-saccadic eye movements. The inability to suppress unwanted glances toward the target or to move eyes away from the target reflects frontal eye field dysfunction and is associated with psychometric evidence of cognitive decline [11, 12]. The neuropsychologic test battery is listed in Table 1. Raw scores were used except for timed tests where a transformed score reflecting performance was used: performance = number test items scored correctly/time (seconds) needed to finish the test. The Delirium Rating Scale (DRS) was used to assess for postoperative delirium. As performance of neuropsychologic tests can be influenced by a patient's mood state and anxiety level, the Hospital Anxiety and Depression Scale (HADS) [13], with subscales evaluating patients for both depression and anxiety, was completed concurrently with the neuropsychologic assessment. The neurocognitive test battery was administered by two evaluators who had been trained under the supervision of a clinical neuropsychologist. Testing occurred in a quiet environment without extraneous distractions. Cognitive decline was defined as a decline of one standard deviation or more in performance on one or more of the neuropsychologic tests.
|
Interventions to treat decreasing rSO 2 included the following: repositioning of the head or perfusion cannulae; increasing arterial carbon dioxide tension, increasing systemic arterial blood pressure, adjusting pump flow rate or anesthetic depth; reduction of temperature; vasodilation; or blood transfusion. To determine the order of intervention, the value farthest from acceptable range was modified first at the anesthesiologist's discretion. Interventions were performed only in the operating room.
The surgical team was comprised of five surgeons and ten anesthesiologists utilizing uniform anesthesia, surgical, and perfusion techniques. Only CABG procedures performed on cardiopulmonary bypass were included for analysis. Standard flow rates of 2.2 L/minute to 2.4 L/minute were utilized to maintain a mean arterial pressure greater than 60 mm Hg, partial pressure of carbon dioxide (arterial) maintained at 40 mm Hg or greater by alpha-stat management, and hematocrit maintained above 22%. Mild hypothermia, antegrade and retrograde crystalloid cardioplegia, and generally a two-clamp technique were utilized.
Baseline and intraoperative data collected for each patient are listed in Table 2. From the intraoperative cerebral oximetry data, an rSO 2 score was calculated by the following formula: rSO 2 score = 50% rSO 2 – current rSO 2 (%) x time (seconds). The rSO 2 score generated is an area under the curve measurement, which accounts for both depth and duration of desaturation below the 50% saturation threshold. Postoperative data, 3 month clinical, and neurocognitive testing data were also collected.
|
) = 0.05 and 80% power, 108 patients per group were needed. To account for possible missing or incomplete data, 133 patients were to be randomized to the intervention and the control group, for a total sample size of 266 patients. Enrollment was terminated at patient 265.
Statistical Analyses
To ensure comparability of the intervention and control groups, demographic and preoperative factors were evaluated. Intraoperative and postoperative factors were also compared in the two study groups. The intervention and control groups were analyzed relative to the presence or absence of postoperative cognitive decline. The study cohort was then divided into cognitive decline and no decline groups and a similar comparison was undertaken.
Categoric data were analyzed using the 
2 or Fisher exact tests. The distribution of all continuous data was examined. For continuous variables whose distributions approximate normality, a t test was used for comparisons. When normality assumptions were not satisfied, the nonparametric Mann-Whitney U test was utilized.
Separate multivariate logistic regression models were used to examine the independent effects of intraoperative intervention and prolonged rSO 2 desaturation on the risk of developing postoperative cognitive decline, cognitive decline at 3 months, and the risk of prolonged hospital stay (length of hospital stay > 6 days). The backward elimination procedure was used to obtain the final models predicting postoperative cognitive decline and prolonged hospital stay. Age, gender, and race were forced into the final models regardless of significance. All clinical factors were examined and factors that showed no further significant improvement to the model fit were eliminated from the final models. Model diagnostics based on the Hosmer-Lemeshow goodness-of-fit method were performed to examine how well the final statistical models fit the data. A multivariate linear regression model was used to assess the independent effect of intraoperative intervention and prolonged rSO 2 desaturation on the average change in cognitive performance at baseline versus the early postoperative phase. Statistical significance was considered at p less than 0.05. All statistical analyses were performed using MINITAB 14.0 (MINITAB Release 14.1; Minitab Inc, State College, PA).
|
Results |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
The definition of prolonged cerebral desaturation was based on a sensitivity analysis of our data. The rSO 2 at 50% represents the average of observed intraoperative cerebral oxygen saturations. Prolonged rSO 2 desaturation was defined as rSO 2 score greater than 3,000%-second below a 50% saturation threshold. Patients who did not desaturate below the 50% threshold had no meaningful incidence of postoperative neurocognitive decline. Analysis revealed that a calculated rSO 2 score of greater than 3,000%-second was associated with increased risk of neurocognitive decline. This value represents the 75th percentile of calculated rSO 2 scores.
Of the 240 patients in the study population, 143 (60%) developed postoperative cognitive decline. No statistically significant difference was found in the incidence of cognitive decline between the intervention (58%) and control (61%) groups. Unadjusted comparisons of demographic, preoperative, intraoperative, and postoperative factors were made between patients with decline and patients without decline (Table 3). The incidence of prolonged rSO 2 desaturation was significantly higher in the cognitive decline group (33%) verdus the no-decline group (20%; p = 0.024). Patients with cognitive decline had significantly longer aortic cross-clamp and bypass times and a higher score on the DRS (p < 0.05). However, after adjusting for demographic and clinical comorbidities aortic cross-clamp and bypass times were not found to be significant in the multivariate logistic regression model for postoperative cognitive decline.
|
|
|
|
|
|
Comment |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
In the present study, the incidence of early postoperative cognitive decline is 60%. This is consistent with previously published reports [1–4]. Cognitive decline was defined as an observed change of one standard deviation or more on a minimum of one neuropsychologic measure. Previous studies examining cognitive function in a CABG population have used various, and differing, criteria to define cognitive decline [1, 4, 19–24]. Mahanna and colleagues [19] applied five different criteria for decline to the same data set and found that the incidence of cognitive decline at 6 weeks after surgery ranged from a low of 1% to a high of 35%, depending on which criterion was used. Our definition of decline is liberal; however, an observed change of one standard deviation not only has statistical meaning but reflects a significant change in performance. A clinical change of this magnitude would cause a patient to be reclassified from normal performance to abnormal performance by an evaluating neurophysiologist [25]. In this interventional study, our goal was to determine if maintenance of cerebral perfusion with the use of cerebral oximetry monitoring had an impact on early postoperative outcomes. While some have cautioned against reliance of data obtained during this "noisy" period for descriptive studies, others agree that assessing the impact of an intervention on short-term outcome would be difficult, if not impossible, to do without these data [10]. We acknowledge that the immediate postoperative period is "noisy" (pain, postoperative medication) in terms of neurocognitive testing, but maintain that these influences were present for both the intervention and control groups and therefore do not adversely affect our findings.
Our results demonstrate that intraoperative rSO 2 desaturation is significantly associated with demonstrable neurocognitive decline in a prospectively randomized CABG population. We found a greater than twofold increase in the odds ratio of developing neurocognitive decline early postoperatively in patients with significant rSO 2 desaturation. Several publications suggest various thresholds below which rSO 2 desaturation is significant; however, they were not prospectively randomized and lacked a control group [7, 26]. One of our objectives was to better identify a meaningful threshold for cerebral oxygen desaturation. In this study cohort, sensitivity analysis revealed that patients who did not desaturate below the 50% threshold had no meaningful incidence of postoperative neurocognitive decline.
Other factors significantly associated with increased risk of cognitive decline include the following: history of smoking; higher score on the postoperative DRS; and absence of preoperative renal insufficiency (creatinine < 2.0 [mg/dL]). Cognitive decline has been associated with smoking and delirium [27, 28]. However, we were surprised to find that preoperative renal insufficiency was associated with a lower risk of cognitive decline. Further analyses revealed that these patients had a significantly lower baseline average performance on several neuropsychologic tests. This finding suggests that these patients are demonstrating a "basement effect" and are unable to decline below baseline levels.
Cerebral oxygen desaturation was also significantly associated with prolonged length of stay. We defined prolonged length of stay as greater than 6 days, which represents the 75th percentile for length of stay in the study population. Patients with prolonged rSO 2 desaturation had a near three times risk of increased length of stay. This claim has previously been made by other authors but not in a randomized population [8]. A more recent study also demonstrated an increased length of stay secondary to an increase in major organ morbidity [9]. This strong association between cerebral oxygen desaturation and length of stay may serve to cost justify cerebral oximetry monitoring.
In the present study, the intervention group trended toward a lower observed rate of cognitive decline, 0.81 odds ratio, but not to a statistically significant degree. Recent studies have unequivocally demonstrated the ability to positively impact cerebral oxygen saturation intraoperatively [9]. Murkin and colleagues [9] were able to normalize cerebral oxygen saturation with one or more maneuvers in 84% of patients in a similarly sized cohort. This was accomplished using the same interventions as in our own study. In the present study, however, the observed desaturation rates in both the control and intervention groups were nearly identical (30% vs 26%, respectively). The observed lack of impact in the intervention group is likely due to poor compliance to the treatment protocol. While the present study demonstrates a clear relationship between cerebral oxygen desaturation and cognitive decline, we were not able to demonstrate as clearly that treatment of rSO 2 will result in better outcomes. Knowing that the relationship between intervention and improved rSO 2 values has already been established we feel confident that future studies will provide further positive reinforcement for aggressive treatment.
Previous studies [1, 22] have demonstrated that early postoperative cognitive decline is, to some degree, transient and that the magnitude of the cognitive deficit decreases with time. Despite this return toward baseline, late neurocognitive decline (5 years) has been associated with the presence of early postoperative deficits [1]. We observed a 29% incidence of cognitive decline at 3 months. Patients with early postoperative cognitive decline had a near twofold increased risk of having cognitive decline at 3 months (OR = 1.86, p = 0.067). There was no difference in the incidence of cognitive decline between the intervention and control groups. The decreased rate of cognitive decline at 3 months observed across both groups is consistent with previous studies [1, 22]. However, by decreasing the incidence of early postoperative cognitive decline through the use of cerebral oximetry monitoring one may have a positive impact on late neurocognitive changes.
As noted above, the major limitation of the current study is the failure to adhere to the intraoperative protocol with regard to the management of cerebral oxygen desaturation. However, this shortcoming only impacts our analysis of assessing endpoints in the intervention versus the control group and has little or no effect on our finding that cerebral oxygen desaturation is significantly associated with adverse outcomes. The strength of the current study is the prospectively randomized design, the quality of the randomization, the adherence to published guidelines on neurocognitive testing, and the uniformity of both surgical technique and neuropsychologic evaluation.
Prolonged intraoperative cerebral desaturation is significantly associated with an increased risk of cognitive decline after CABG, as well as increased length of hospital stay. Intraoperative management of cerebral desaturation may result in decreased postoperative cognitive decline and less frequent prolonged length of hospital stay after CABG. We believe that the present findings support validation and further investigation in a multicenter trial.
|
Discussion |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
What changes in processes of care have you made in your group as a consequence of what you've learned about the causes hypoperfusion that have caused the desaturation? Have you considered monitoring gas and microparticulate emboli, which potentially effect cognitive decline as well?
DR SLATER: I'll take your second question first. We do not monitor microemboli routinely.
In terms of what we learned, it has been an ongoing process. For instance, we found that in augmenting blood pressure, Levophed had a more positive effect on oxygen saturation than Neo-Synephrine. So we gradually gravitated towards use of Levophed over Neo-Synephrine in the operating room. In addition, we've been running our PCO2s quite a bit higher than we had previously as this tended to have a positive effect on cerebral oxygen saturation. We continue to monitor various intraoperative treatments to determine their impact on cerebral oxygen saturation.
DR JOHN ALEXANDER (Evanston, IL): I practiced in your community at Hackensack University Medical Center for a period of 7 years. We used this device for nearly the entire period of time that I was there. I am absolutely convinced that the conclusions that you have presented today are exactly right. I will tell you some of the lessons that I thought we learned, and I would be interested in your comments.
You mentioned CO2. I think CO2 is a very important part of this puzzle. The anesthesiologist puts the patient to sleep, intubates him, sets the oxygen at 100%, and the ventilator at a minute ventilation that is appropriate for an awake patient. When the metabolic activity of the patient decreases as he is put to sleep and gets cold, they don't need nearly as much ventilation. If you watch the PCO 2 [partial pressure of carbon dioxide] during this time period, it will go down as the ventilator blows off CO2.
In some patients who are susceptible to a neurologic event, when they get lower PCO 2s, they vasoconstrict in their brain, and you'll see their brain numbers come down. So your attention to PCO 2 is really very important. Carrying the PCO 2 higher is critically important in some patients. It was one of the physiologic lessons that I have learned from watching the "brain number." I learned about this in medical school, but it never was of any practical importance to me until this machine taught me how important it can be in some patients.
The other issue is the use of blood. We have all tried not to use blood for a variety of reasons. But in the susceptible patient, and you never know who they will be, when the hematocrit gets to around 20, and the "brain number" decreases, we found that the very best thing to do to correct the brain number was to give blood. We ultimately wound up using a little bit more blood, but I think we had a much better neurologic outcome in our patients.
I am also convinced that some of the soft positive effects of off-pump surgery, in the sense of the patients waking up quicker and looking better, is a blood effect. Off-pump patients never get nearly as hemodiluted as the on-pump patients. I suspect that some of the soft improvements after "off pump" CABG [coronary artery bypass grafting] are the results of this blood effect. The small prime volume pump paper that we saw earlier in this session is also a higher hematocrit story. It seems to me that all of these things are tied together. In my opinion this device is the best real time way to measure the end organ effect of adequate perfusion.
The detrimental secondary effects that are associated with cardiopulmonary bypass are minimized if you keep the brain number good. A good brain number means the body is happy and you'll have less renal failure and less complicating perioperative events. That's why these people have a better length of stay. They're not as damaged by the heart-lung machine.
I guess after having said that, what are you specifically doing with PCO 2 and have you found that blood is important?
DR SLATER: We're certainly, as I said in the response to the last question, keeping PCO 2s more physiologic, typically in the 40s. We did not do this previously. I think it's important to do that not only during the surgery, but immediately after as well when the patients are still cool and systemically underperfused.
It was difficult to mandate blood transfusion during the study based on a value (cerebral oxygen saturation) that in our minds was as of yet unproven. That is why we chose to test two hypotheses. A positive finding on first hypothesis, that is, that oxygen desaturation increases risk of neurocognitive decline, would justify more aggressive treatment, including transfusion. I agree with you that maintaining hemoglobin is important with regards to cerebral oxygen saturation.
DR ROBERT J. RIZZO (Boston, MA): I agree with Dr Alexander's comments. We've used this as well for all of our cases for the last 400 or 500 at Cape Cod Hospital. I've noticed that by continually monitoring this, you can find all kinds of things that you didn't even think about that were related to brain desaturation, including when you go on bypass, particularly if you go on without retrograde autologous prime, you'll see a sudden drop in the arterial saturations; with autologous prime, you don't see that.
With retraction of the heart, you may see desaturation. Even with sizing of grafts on bypass you'll notice that desaturation occurs. All of these little things, I don't know if you noticed that as well, but you start getting all this feedback about things happening to the brain that you didn't even think about were happening when you're working on the heart. So I find it pretty helpful technology.
I've noticed since using it we are using a little more blood. I think we're getting a little better neuro-outcomes with using more blood.
And then the other area is the particulate emboli, including air, which we have more in the heart than we think of when we perform a CABG on bypass. And we're a little more aggressive now at the end of the case deairing the aorta out the graft proximals, and I think that also helps us have less desaturation to the brain coming off. Have you noticed any of those things as well in your experience?
DR SLATER: We have made similar observations, and do our best as the operating surgeon, to minimize those actions that depress cerebral oxygenation the most. We haven't specifically monitored for or addressed the emoboli issue to date.
DR RIZZO: Thanks.
DR SLATER: In addition to my comments to the last two questions I would also add that in January's Anesthesia there is a publication that evaluates intraoperative cerebral oxygen monitoring and its impact on organs other than the brain. This study suggests that by monitoring and treating cerebral oxygen desaturation one observes less other organ dysfunction such as renal and pulmonary insufficiency. This implies that management of cerebral oxygen saturation may have an even larger impact than initially thought.
DR JOSEPH C. CLEVELAND (Denver, CO): It's a phenomenal study. And in a way this is a little bit off the beaten track, because it's unfocused, but just comparing this again to an age-matched group of 70-year olds undergoing lobectomies, which you may have access to, have you measured them and just seen if they have episodes of desaturation? Obviously, all of us here wouldn't argue that. I think that the thesis that you should measure this, I think I agree with totally, but have you looked at that group just as a control group, if you will?
DR SLATER: Our study is restricted to a coronary surgery population. One could imagine potential benefit for any group of patients undergoing a major operation.
The literature available at the time we conceived the study was limited to observation studies only. We identified a real need for a prospective, randomized trial. Because we are heart surgeons that was the population we chose. Within that group, primary CABG patients represented the most uniform group. There is no reason not to believe that other groups of patients would not benefit as much if not more than our studied population.
Hopefully our study along with the other recent publications that I mentioned will spur additional work around these issues.
|
Acknowledgments |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
|
References |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| 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 |
