In this exploratory study, we found evidence of depression in 15.6% of patients before surgery and in 13.4% of patients after surgery. Of patients with preoperative depression, 50% remained depressed one month after surgery while 6.6% of patients without preoperative depression demonstrated new evidence of depression after surgery. Preoperative depression was not associated with change from baseline in cognitive performance postoperatively. In contrast, depression detected one month after surgery was associated with worse postoperative cognitive performance in the domains of attention, memory, verbal fluency, processing speed, and fine motor function. In contrast to preoperative depression, those with depression after surgery were more likely to demonstrate the composite neurocognitive dysfunction outcome one month after surgery. Preoperative depression was associated with postoperative impairment in several quality of life domains and with higher anxiety. Depression one month after surgery was associated with impairment on all but one of the quality of life domains and with measures of anxiety.
There is a biologically plausible reason to expect that depression might impair neurocognitive recovery from surgery. Monoamine neurotransmitter imbalance in the central nervous system is one proposed mechanism of depression providing the rationale for treatment with drugs that increase their brain concentrations (e.g., tricyclic antidepressants, selective serotonin reuptake inhibitors, monoamine oxidase inhibitors). Maladaptive inflammatory responses to various stressors leading to elevation in pro-inflammatory cytokines such as interleukin 1β, interleukin 6, and tumor necrosis factor α is another proposed mechanism of depression.[28,29] Altered neurotransmission and neuroinflammation are both implicated as putative mechanisms of perioperative neurocognitive disorders supporting the potential for a common underlying mechanism that overlaps with depression.[11,30]
In community-dwelling cohorts cognitive impairment has been shown to accompany depression. Neurocognitive dysfunction is an important source of patient morbidity that can complicate recovery from cardiac surgery.[11,32] In contrast to non-surgical subjects, there are little data, on whether depression is associated with postoperative neurocognitive dysfunction. In this study we did not find a link between preoperative depression and neurocognitive performance one month after surgery. In an analysis of data collected from 123 patients undergoing CABG surgery enrolled in the Neuropsychiatric Outcomes After Heart Surgery (NOAH) study, preoperative depression defined with the Hamilton Depression Rating Scale was detected in 9.75% of patients.[33,34] In contrast to our finding, preoperative depression was associated with decline in cognitive performance one month after surgery measured with the clinical dementia rating scale sum of boxes (CDR-SB) score. The CDR-SB inventory is a composite measure of cognitive and functional performance used for dementia rating. Postoperative depression screening results were not reported. Meta-analysis of data from non-surgical cohort studies have demonstrated that deficits in executive function, memory, verbal fluency, and attention are more common in depressed than non-depressed subjects. Not unexpectedly, our data and that from secondary analysis of the NOAH study show that there is overlap in the cognitive domains impaired with depression in patients undergoing cardiac surgery as observed in population-based studies. Our data, though, suggests that depression after surgery might additionally impact processing speed and fine motor speed after surgery that could affect functional outcome. Regardless, mild depressive symptoms can fluctuate and this may explain, in part, the absence of depression after surgery in 50% of patients classified as having preoperative depression. Other patients without depression before surgery could be classified as depressed postoperatively. Limiting depression screening to the preoperative period only, thus, may not detect clinically impactful depression after surgery.
Our findings are in contrast to prior work by our group and that of others showing that pre-existing depression increases the risk for perioperative neurocognitive disorders manifested as delirium after cardiac and non-cardiac surgery.[13–15,35,36] In our prior report, we found that a history of depression was associated with a three-fold (95% confidence interval, 1.3 to 7.0, p = 0.011) increased risk for delirium after cardiac surgery. In the present study, we objectively determined the presence of depressive symptoms rather than classifying patients as having depression based on medical history. This more sensitive measure identified sub-clinical forms of depression since the majority of depressive symptoms were mild. Regardless, whether perioperative depression contributes to postoperative neurocognitive dysfunction, or vice versa, can not be ascertained from this retrospective analysis. Future prospective investigations are necessary to determine if indeed depression is a potentially modifiable risk for postoperative delirium or neurocognitive dysfunction.
One explanation for our findings of reduced cognitive performance in depressed patients after surgery is that affected patients may show reduced effort during neuropsychological testing which can be rigorous. Meta-analysis from population-based studies, though, show that cognitive impairment is a component of depression and not an epiphenomenon that is secondary to low mood. Our data raises the possibility that targeting treatment of postoperative depression might provide a strategy for improving postoperative neurocognitive function. Treatment with antidepressant drugs alone, while safe in patients with cardiac disease, takes weeks to show benefit and then may have sub-optimal efficacy.[37,38] Members of our team have previously reported results from a randomized study of 123 patients with DSM-IV criteria for major or minor depression within 1 year of cardiac surgery demonstrating the benefits of cognitive behavioral therapy or stress management counseling compared to usual care. Remission of depression occurred by 3 months of therapy in a greater proportion of patients in the cognitive behavior therapy (71%) and the supportive stress management (57%) groups than in the usual care group (33%, p = 0.002). Of note is that that study did not focus on neuropsychological end-points detected in the early months after cardiac surgery. Others have shown benefit from an 8-month telephone delivered collaborative care intervention for improving mental health quality of life, physical function and depressive symptoms after CABG surgery. In that study, patients were screened and enrolled before postoperative hospital discharge using the Physical Health Questionnaire-2. A nurse manager performed active follow-up of patients adhering to evidence-based treatment protocols, patient education about their illness, patient history and preferences for treatments while actively involving their primary care physician and transfer of care to a mental health clinician if the patients did not respond to therapy.[39,40]
Depression at the time of cardiac surgery is known to be associated with postoperative complications, slower physical and emotional recovery, impaired quality of life and mortality.[3,41–43] Our observation of impairment on some quality of life measures and higher anxiety in depressed versus non-depressed patients, thus, confirms data from other studies. The small number of patients in our study, though, limits the interpretation of the relationship between preoperative depression and risk for major complications. History of overt and covert cerebral ischemic injury as well as narrowing of cerebral arterioles ("small vessel disease") is known to be associated with depression in the general population.[44,45] In our study patients with depression had a higher prevalence of known stroke before surgery but there was no indication of any difference in the prevalence of cerebral hemodynamic abnormalities in depressed compared with non-depressed patients. That is, the lower limit of cerebral blood flow autoregulation, a parameter that is higher in patients with prior stroke, was not different between these groups. Further, the frequency of impaired cerebral blood flow autoregulation during cardiopulmonary bypass was not different in depressed and non-depressed patients. In a prior study, we found that patients with brain magnetic resonance imaging evidence of cerebral small arteriolar disease had a higher frequency of impaired cerebral autoregulation during cardiopulmonary bypass.
In this study, we did not find a relationship between preoperative depression and clinically detected delirium in contrast to our prior studies.[13,14] Since many episodes of postoperative delirium are manifest as the hypoactive sub-type, failure to perform a structured delirium assessment such as with the Confusion Assessment Method likely led to our failure to account for all episodes of delirium. Our results are likely confounded by the small number of patients with delirium. Patients in this study were enrolled in a prospectively randomized trial evaluating whether basing mean arterial pressure targets during cardiopulmonary bypass based on individualized cerebral autoregulation data reduces the frequency of neurological complications compared with usual care. In that study, we found that patients in the autoregulation intervention arm had a lower frequency of clinically detected delirium and improved performance on test of memory one month after surgery from baseline. In the current depression study, we adjusted analysis for treatment arm of the parent study minimizing any influence of the autoregulation blood pressure management intervention may have on our observations. Another limitation to this study is missing cognitive data. Missing cognitive data usually involved a small number of test and is often due to the patient's request to not continue neuropsychological testing which can be challenging especially in those with cognitive impairment. We did not a priori impute missing neuropsychological test scores, however, we did perform imputations in a sensitivity analysis. The latter analysis did not markedly change our findings. Although we did not utilize a formal diagnostic schedule administered by a mental health clinician to validate a psychiatric diagnosis of Major Depressive Disorder, presence of increased depressive symptoms determined by using questionnaires such as the Beck Depression Inventory in non-psychiatric settings have been shown to be clinically significant in improving outcomes and harm reduction.[49,50]
In conclusion, this exploratory analysis of data collected from a randomized study suggests that preoperative depression is largely not associated with perioperative neurocognitive dysfunction but it may impact postoperative quality of life. In contrast, depression detected one month after cardiac surgery may be associated with postoperative impairment in attention, verbal fluency, processing speed, and fine motor function as well as altered quality of life.
The authors acknowledge with appreciation for the support provided by our surgical, anesthesiology, nursing, an perfusion colleagues for the implementation of this study.
This study was funded in part by NIH RO1HL092259 (Charles W. Hogue, MD, PI); NIH K76 AG057020 (Charles H. Brown, IV, MD, MPH).
Availability of data and materials
Per the charter of the original grant submission, the database from this study will be made available by the corresponding author upon a reasonable request with a clear statement of purpose and aims.
Ethics approval and consent to participate
The parent study was approved by the Institutional Review Boards of The Johns Hopkins Medical Institutions, Baltimore, MD and Northwestern University, Chicago, IL. Patients provided written informed consent to participate in the study.
All methods were carried out in accordance with relevant guidelines and regulation.
Consent for publication
There is no personal identifying information for the enrolled patients included in this publication. The consent for publication by the participants of this study is not applicable.
BMC Anesthesiol. 2022;22(157) © 2022 BioMed Central, Ltd.