Postoperative cognitive decrease (POCD) represents the most frequent complication in modern cardiac surgery. The application of easily assessable surrogate parameters that predict long-term POCD at early time points is tempting. The aim of the present study was to analyze the predictive value of cerebral biomarkers, diffusion-weighted magnetic resonance imaging (DWI), and cognitive bedside testing after coronary artery bypass grafting (CABG). From 106 patients who underwent elective CABG, blood samples were drawn for the measurement of protein S100B and neuron-specific enolase release at baseline, at the end of surgery, and 48 hours afterward. Cerebral DWI was carried out before and 2 to 4 days after surgery. Cognitive functioning was assessed before, 2 to 4 days (bedside testing) after, and 3 months after CABG. On DWI, lesions were detected in 15.1% of patients. Biomarker levels and the presence of acute ischemic lesions on DWI were not associated with long-term POCD. Early POCD was correlated with 3-month POCD (r = 0.46, p <0.001). Ninety-one percent of patients who had shown moderate to severe POCD (<−1.5 z scores) in the early phase still had decreased memory functioning at 3 months compared to baseline (likelihood ratio 5.23). Early POCD was asserted as the only predictor for long-term POCD in a stepwise multiple linear regression model (R 2 = 0.20, p <0.001), excluding age, length of surgery, aortic clamping and cardiopulmonary bypass duration, the number of anastomoses, and postoperative neuron-specific enolase and S100B levels. In conclusion, the results show that in contrast to biomarkers, DWI, age, or intraoperative variables, early neuropsychological bedside testing predicts long-term POCD after CABG with acceptable accuracy.
Postoperative cognitive decrease (POCD) represents the most frequent complication in modern cardiac surgery. The application of easily assessable surrogate parameters that predict long-term POCD at early time points is tempting. Besides age and intraoperative characteristics such as the number of anastomoses or procedure length, other parameters have been discussed as potential candidates. Neuron-specific enolase (NSE) and astroglial protein S100B correlate with the extent of substantial brain damage and POCD. However, only a limited number of studies have included neuropsychological assessment for later than immediate (>1 month) outcomes and have reported conflicting results on the predictive value of biomarkers in this regard ( Table 1 ). On diffusion-weighted magnetic resonance imaging (DWI), focal hyperintense spots, suggestive of small embolic ischemic lesions, can be detected frequently after heart surgery. Because an incidence of DWI lesions up to 51% has been reported, the hypothesis that these lesions represent the morphologic correlate for POCD appears striking. Correlative analyses between DWI lesions and long-term POCD were performed only in a few smaller studies that provided conflicting results ( Table 1 ). Finally, neuropsychological bedside testing in the acute phase after coronary artery bypass grafting (CABG) might be influenced by aftereffects of major surgery, such as pain, postoperative confusion, or analgesia. However, a variety of studies have indicated that early POCD predicts outcomes after months or years. The present study was conducted to analyze the association of all these parameters with long-term POCD in a large and homogenous cohort of patients who underwent CABG.
| Study | Sample Size/Type of Surgery | Markers | Time of Postoperative Blood Sampling | Time of Postoperative Neuropsychological Assessment | Postoperative DWI | Conclusions | Prediction of Long-Term Neuropsychological Deterioration | |
|---|---|---|---|---|---|---|---|---|
| NSE or S100B | DWI Lesions | |||||||
| Ramlawi et al | 40/CABG and OHS † | S100B, NSE, τ | 6 hours, 4 days | 4 days, 3 months | No |
| No | — |
| Jonsson et al | 110/CABG | S100B | 1–10, 15, 24, and 48 hours | 14 days, 2 months | No |
| No | — |
| Westaby et al | 100/CABG | S100B | End of CPB, 8 hours | Before discharge, 3 months | No |
| No | — |
| Herrmann et al | 55/CABG; 19/OHS | S100B, NSE | 1, 6, 20, and 30 hours | 3 days, 8 days, and 6 months | No |
| No | — |
| Kilminster et al | 109/CABG; 21/OHS | S100B | 0.5, 1, 1.5, 2, 3, and 5 hours after CPB start | 8 weeks | No |
| Yes | — |
| Lloyd et al | 30/CABG | S100B | During and 4, 12, and 24 hours after CPB | 12 weeks | No |
| No | — |
| Barber et al | 40/OHS | No | — | 5 days, 6 weeks | Yes |
| — | Yes |
| Cook et al | 25/CABG; 25/OHS ⁎ | No | — | 4–6 weeks | Yes |
| — | No |
| Knipp et al | 29/CABG | No | — | 3 months | Yes |
| — | No |
| Knipp et al | 30/OHS | No | — | 5 days, 4 months | Yes |
| — | No |
| Knipp et al | 39/CABG | No | — | At discharge, 3 months, and 3 years | Yes |
| — | No |
Methods
Patients scheduled for elective CABG were prospectively enrolled. The following exclusion criteria were applied: age <18 years, contraindication to magnetic resonance imaging, additional surgery (i.e., heart valve replacement), previous surgery using cardiopulmonary bypass (CPB), and the detection of a significant (>70%) stenosis of the brain-supplying arteries on preoperative magnetic resonance angiography. The study was approved by the ethics committee of Justus-Liebig University Giessen. All patients gave informed consent.
Cognitive examination was performed by an experienced neuropsychologist, using a battery of well-established tests ( Table 2 ). In all domains, parallel test forms were used postoperatively, when available. At 2 to 4 days, an abbreviated form of the test battery was administered. It consisted of Syndrom Kurztest (SKT), which is a rating scale to assess functions of memory and attention. In the broader battery at baseline and 3 months, using other tests in addition to SKT, cognitive function was tested more thoroughly (e.g., to assess different aspects of declarative memory). All raw scores of the psychological tests were converted into z scores as described by Gerriets et al. Higher z scores always represent better test performance. Results were shown to vary greatly dependent on the criteria of POCD; for that reason, POCD was not defined by a categorical definition, which might be flawed by subjectivity and arbitrariness. Instead, POCD was described as a statistically significant change from the baseline performance level on a major test or domain, using parametric statistics with α correction.
