Postoperative cognitive decline (POCD) is a well-recognized phenomenon after cardiac surgery. Cerebral embolization, systemic inflammatory response, and cerebral hypoperfusion are usually implicated as likely etiologic factors for short-term POCD after cardiac surgery. The incidence of short-term POCD varies among studies, but the reported rates are 46-53% at discharge1 , 2 and 15-36% at six weeks1 , 3 , 4 after coronary artery bypass graft (CABG) surgery. The presence of early POCD measured at hospital discharge has been shown to be a significant predictor of long-term cognitive dysfunction five years after surgery.1 , 2 However, it has also been suggested that the presence of cerebrovascular disease at baseline and the subsequent development of new cerebral infarcts in the elderly patient may play an important role in late cognitive decline irrespective of surgery and cardiopulmonary bypass (CPB).5 - 8

Recently, we showed that processing of shed blood with a continuous-flow cell saver during CPB resulted in a clinically significant reduction in POCD six weeks after surgery.4 The reduced cognitive decline in the cell saver group was likely attributed to the lower lipid cerebral embolic load and the modified systemic inflammatory response with the application of cell saver.

The purpose of the current study was to determine if the early benefit of reduced POCD was sustained in the same patient population at the one-year follow-up.

Methods

After approval by the University Health Network Research Ethics Board (10th Floor, Room 1056, 700 University Ave., Toronto, ON, M5G 1Z5, August 16, 2001, REB 01-0420-B), an informed consent was obtained from patients > 60 yr of age who were scheduled for elective CABG surgery. Exclusion criteria included a surgical procedure in addition to CABG, redo CABG surgery, emergent surgery, severe kidney or liver disease (creatinine > 133 mg·dL−1 and bilirubin > 2 mg·dL−1), symptomatic cerebrovascular disease, history of stroke, transient ischemic attacks, or atrial fibrillation. We also excluded patients who were unable to complete the preoperative assessment and those who could not speak English. Patients were randomly allocated to either a cell saver or a control (cardiotomy suction) group according to a computer-generated randomization code in blocks of a predetermined size. Patients were followed for up to one year after surgery.

Perioperative management strategies were described in detail in the original publication.4 Briefly, the continuous-flow cell saver (Fresenius Corporation, Concord, CA, USA) was used to process shed blood before returning it back to the patient, and cardiotomy suction was used in a standard closed venous reservoir where cardiotomy blood was collected and reinfused through the arterial circuit back to the patient. Both the cell saver and the cardiotomy suction were used during the same time periods, i.e., from the time full heparinization (ACT > 400 sec) was established to after the initial dose of protamine was given.

Neuropsychological testing

A trained psychometrist blinded to the treatment arm assignment conducted the neuropsychological testing one week before (baseline) surgery as well as six weeks and one year after surgery. The battery of tests included 12 tests. Ten main variables were chosen a priori to be used in the analyses: 1) Rey Auditory Verbal Learning Test, 2) Rey Visual Design Learning Test, 3) Halstead-Reitan Trail Making Tests Parts A and B (Trails B - Trails A), 4) Grooved Pegboard Test, 5) Wechsler Memory Scale (WMS) Digit Span Forward, 6) WMS Digit Span Backward, 7) WMS Spatial Span Forward, 8) WMS Spatial Span Backward, 9) Choice and Simple Reaction Time Tests (Choice Reaction Time - Simple Reaction Time), and 10) Verbal Fluency Test. If a lower score showed improved performance, the directional data were reversed so that all improvements gave positive change scores.

The same analysis of the neuropsychological testing was applied as previously described.4 In brief, to estimate the change in performance from baseline to one year after surgery, the raw scores for each test were converted to Z-scores. A Z-score was calculated for each main variable in each patient by subtracting the preoperative score from the postoperative score and dividing the difference by the preoperative standard deviation of that variable.9 Subsequently, a combined Z-score was calculated by the summation of all of the individual test scores. Patients with a negative score of < 1 were considered to have POCD. The primary outcome was the dichotomous cognitive deterioration variable based on the combined cognitive score. Tests not completed were treated as omissions and not as failures.

Sample size justification and statistical analysis

Sample size calculations were based on the prevalence of POCD at six weeks after surgery and were described in detail in the original publication.4

Comparability of both groups was tested with the use of Chi square statistics or Fisher’s exact test on qualitative variables, as appropriate, and the Student’s t test on quantitative variables. For the primary analysis of dichotomous cognitive outcome, a sum of the individual Z-scores (a combined Z-score) of the ten main variables was compared between the two groups with a Chi square test. Confidence intervals (CI) for proportions were calculated at 95%. A P value < 0.05 was considered significant. All analyses were performed on an intention-to-treat basis. Statistical analysis was conducted with the use of SPSS® computer software (SPSS Inc., Chicago, IL, USA).

Results

In the original trial, 226 patients (cell saver group, n = 112; control group, n = 114) underwent baseline neuropsychological testing. Ninety-nine patients in each group were evaluated for the presence of POCD at six weeks after surgery. The observed proportion of patients with POCD was six of 99 (6.1%) patients in the cell saver group and 15 of 99 (15.2%) patients in the control group, difference, 9.1% (95% CI, 0.4 to 18.1). Eighty-four of 112 (75%) patients in the cell saver group and 86 of 114 (75%) controls returned for one-year neuropsychological testing (Fig. 1). One year after surgery, postoperative cognitive decline was present in 16 of 84 (19%) patients in the cell saver group (95% CI, 10.8 to 27.2) and 15 of 86 (17.4%) patients in the control group (95% CI, 9.6 to 25.2) (P = 0.786) (difference, -1.6%; 95% CI, -13.3 to 10.1) (Fig. 2). Both groups were similar with respect to demographic data and surgical characteristics (Table 1).

Fig. 1
figure 1

Schematic of neuropsychological assessment during the one-year follow-up after coronary artery bypass graft surgery

Full size image
Fig. 2
figure 2

The rates of postoperative cognitive decline in the cell saver and cardiotomy suction (control) groups at six weeks and one year after coronary artery bypass graft surgery

Full size image
Table 1 Demographic variables and surgical characteristics
Full size table

Twenty-one patients in the cell saver group and 17 patients in the control group who completed the neuropsychological testing at six weeks did not return for the one-year follow-up. Six patients in the cell saver group and four patients in the control group who did not complete the assessment at six weeks returned for the one-year testing. Six of the 15 patients in the control group with POCD at six weeks still had the impairment at one year and five did not; four were lost to follow-up. Three of the six patients in the cell saver group with POCD at six weeks still had the impairment at one year, two did not, and one was not tested. Thirteen (15.4%) and nine (10.5%) patients in the cell saver and control groups, respectively, developed new POCD which was not evident at the six-week follow-up. The raw neuropsychological test scores are shown in Table 2.

Table 2 The raw scores of the baseline and one-year follow-up neuropsychological tests
Full size table

Discussion

The current study showed that the short-term (six weeks) preservation of cognitive function4 in elderly patients using the cell saver management strategy during CPB did not translate into a long-term benefit one year after CABG surgery. Although the lack of long-term preservation of cognitive function after cardiac surgery is disappointing, these findings clearly raise a number of questions regarding the etiology and pathophysiology of POCD. First, do early and late POCDs have different etiologies? Second, is cardiac surgery a form of stressor that identifies patients who would have developed POCD regardless of the surgical intervention? Third, are the perioperative efforts to reduce early POCD worthwhile if the long-term benefits are not sustainable?

It is important to emphasize that the majority of patients who had POCD at the one-year follow-up developed new cognitive dysfunction that was not evident at six weeks after surgery. These findings are supportive of our current contention that different etiologies are responsible for early and late POCD. However, an alternative hypothesis is also plausible, i.e., that the pathophysiologic neurologic sequelae processes (e.g., an apoptotic change initiated by perioperative insults and inflammation) that may have been present at surgery manifested themselves only at the one-year follow-up.

It is well known that the signs of systemic atherosclerotic disease are associated with global cognitive decline in elderly patients.10 , 11 Leary et al.7 reported that 3-5% of the general population in their 70 s are likely to have a new brain infarct each year as detected by magnetic resonance imaging (MRI) scans. Indeed, a link between brain infarcts identified by MRI scans and cognitive decline has been established in large-scale population-based longitudinal studies.12 , 13 A recent editorial by Hammon and Stump8 estimated that the rate of first-time one-year neurological events in the CABG patient population would range from 10-15%. Furthermore, evidence is emerging that preexisting cerebrovascular disease may play a pivotal role in both short- and long-term POCD.14 - 17 Consequently, it is not surprising that the rates of new brain infarcts in previous studies and the new onset of POCD in the current report were similar at the one-year follow-up.

Age and the baseline cognitive index are generally considered the most powerful predictors of both short- and long-term cognitive decline in cardiac as well as non-cardiac surgical populations.1 , 3 , 18 Importantly, several groups of investigators have found similar rates of long-term cognitive decline in age-adjusted non-surgical controls. In the late 1990 s, Hlatky et al. showed that long-term cognitive function was similar after CABG surgery and coronary angioplasty in the majority of patients.19 The International Study of Postoperative Cognitive Dysfunction (ISPOCD) group20 showed that POCD was present in 10.4% of elderly patients one to two years after non-cardiac surgery. This rate of cognitive decline was almost identical (10.6%) to the age-adjusted non-surgical controls tested at the same time interval. At one, three, and six-year follow-ups, Selnes et al. observed similar rates of late cognitive decline in patients undergoing coronary revascularization procedures and those who were of similar age with coronary artery disease who had not undergone CABG surgery.17 , 21 After adjusting for age, sex, education, and baseline comorbidity, there was no significant difference between the rates of cognitive decline in patients after CABG surgery and control subjects without coronary artery disease at five-year follow-up.22 Likewise, using healthy and non-surgical control groups, Sweet et al. could not show a higher late cognitive decline in CABG patients.23

Similarities in late cognitive decline between patients with and without surgery may suggest that age may simply be a surrogate measure for the preexisting cerebral vascular disease which is considered a major contributor to POCD. Even more so, late cognitive decline after CABG surgery may not be specific to employment of CPB, but rather, it may be related to the presence of cerebrovascular disease. However, it seems likely that certain subgroups of patients may have a more favourable course of cognitive recovery depending on the effects that their baseline cognitive function and education levels may have on aging their arterial and immune systems. In the current study, it is significant that the POCD, which was present in one-third of the patients at six weeks, had resolved at the one-year follow-up.

While short-term POCD after cardiac surgery can likely be attributed to macro- and micro-embolism, cerebral hypoperfusion, and systemic inflammatory response, the presence of progressive cerebrovascular disease is likely a contributor to (along with other factors) long-term POCD. Although it appears that etiologies for short- and long-term cognitive impairment differ, the fact remains that one-half of the patients who were found to have POCD at six weeks still had the impairment at one year. This finding supports our contention that implementation of strategies aimed to reduce short-term cognitive impairment might also benefit patients in the long term. Perception of general health varies directly with cognitive functioning. Consequences of cognitive decline in the elderly after non-cardiac surgery include poorer quality of life, risk of leaving the labour market prematurely, dependency on social transfer payments, and increased mortality.18 , 24 Similarly to non-cardiac surgery, it has been shown that patients with POCD had poorer quality of life when assessed at one year and five years after otherwise successful cardiac surgery.25 , 26 Lower five-year overall cognitive function scores were associated with lower general health and a less productive working status. These findings have significant social and financial implications. Unfortunately, there is no known treatment of POCD once it occurs. Therefore, development of reliable and feasible preventive strategies for both immediate and delayed postoperative neurocognitive complications is of paramount importance.

One of the common limitations of longitudinal cohort studies is patient dropout rates on the follow-up assessments. In the current study, a similar number of patients in the cell saver and control groups (21% vs 17%, respectively) failed to return for the one-year follow-up. However, there was no difference between groups with respect to demographic data or surgical characteristics. Furthermore, the focus of the current study was on an elderly patient population undergoing primary CABG surgery. Consequently, our findings should be applied with caution in younger patient populations undergoing more complex cardiac surgical procedures.

In conclusion, the current study showed that short-term preservation of cognitive function in elderly patients using the cell saver management strategy during CPB did not translate into a long-term benefit one year after CABG surgery. The presence of progressing cerebrovascular disease may be a likely contributing factor to long-term POCD.