Introduction
Cardiovascular risk factors are known to be associated with cognitive decline [
1] and dementia [
2]. Conversely, low levels of cognitive function have long been associated with greater all-cause [
3] and cardiac mortalities [
4]. In particular, much attention has been paid to cognitive function and coronary artery bypass graft (CABG) surgery, with estimates suggesting that cognitive impairment is present in approximately 35 % of patients prior to surgery [
5] and in between 22.5 and 42 % of CABG patients following surgery [
6,
7]. The mechanisms linking cognitive impairment and cardiac disease are not clearly understood but are likely to include both inflammatory and neuroendocrine pathways.
Whilst mixed evidence exists, inflammation is thought to be a common causal factor in both coronary heart disease (CHD) [
8,
9] and cognitive decline [
10]. CABG surgery is associated with an acute inflammatory response. The extent of the inflammatory response is thought largely to reflect the amount of trauma derived from the surgical procedure itself and is associated with a host of clinical outcomes [
11,
12]. Likewise, heightened levels of C-reactive protein (CRP) have been associated cross sectionally with mild cognitive impairment [
13] and there is some evidence for a longitudinal effect, with both CRP and interleukin (IL)-6 being associated with worsening of cognitive function in some population studies [
14‐
16]. However, others have not corroborated such findings [
17]. Therefore, more work is needed to study the direction of this effect. As yet, the relationship between cognition and inflammation has not been studied in a CABG population.
The hormone cortisol is also posited to be a potential mechanism linking CHD and cognitive function. For example, in the cardiac literature, elevated cortisol has been linked to cardiac risk factors such as hypercholesterolemia, diabetes, abdominal obesity and hypertension by some [
18,
19], but not all [
20], authors. There are limited studies investigating cortisol in CABG surgery patients, although there is some evidence that there is heightened cortisol output in the post-operative period [
21,
22]. Cortisol has also been linked to poorer cognitive function in observational and experimental studies [
23]. Associations between the area under the curve (AUC) of cortisol and worse performance in six cognitive domains have been observed in a large cohort study [
24]. Higher morning cortisol has been associated with mild cognitive impairment and a lower global cognitive state in older people [
25]. In addition, a prospective association has been observed, such that high morning cortisol and a flatter slope have been associated with poorer cognitive performance 4 years later [
26]. To date, only one study has investigated cognitive function and cortisol in a CABG population. Mu and colleagues [
27] found that CABG patients with higher serum cortisol levels in the first post-operative morning had increased risk of cognitive dysfunction 7 days after CABG surgery. However, little is known about the impact of pre-existing cognitive state on later neuroendocrine adaptation to CABG surgery. It is possible that the neurocognitive changes resulting in impairment to memory and concentration cause systemic changes to the immune system via dysregulation of the hypothalamic-pituitary-adrenal (HPA) axis. Such changes would negatively affect the body’s ability to self-regulate following a physical trauma such as CABG surgery.
One way in which to explore the relationship between cognition and these biological mechanisms is to study the effect of pre-surgical cognitive performance on inflammatory and cortisol responses to CABG surgery. This allows the temporal relationship between cognition and biomarkers to be investigated. Therefore, the aim of the present study was to assess the prospective association between cognitive function prior to CABG surgery and subsequent inflammatory and neuroendocrine responses to surgery. Specifically, we hypothesised that poorer cognitive function would be associated with greater inflammatory responses to surgery, as measured by IL-6 and CRP, and increased cortisol output in the months following surgery, independent of mood factors and other clinical and sociodemographic confounders.
Discussion
The results of these analyses show that low cognitive functioning prior to surgery is predictive of impaired biological responses following CABG, including a heightened IL-6 response during the in-hospital stay and greater cortisol output up to 2 months after surgery. These results were supported in analyses using quintile MoCA scores to predict IL-6 and cortisol AUC responses, indicating a dose–response relationship. Pre-operative cognitive functioning was not associated with hs-CRP following surgery. This is the first study to our knowledge that has examined the prospective association between pre-operative cognitive functioning and biological responses to CABG surgery.
We found that approximately 40 % of participants prior to CABG surgery had low cognitive functioning as measured on the MoCA. This is in line with data from an Australian study which reported cognitive impairment in 35 % of CABG patients prior to surgery, as compared with healthy controls [
5]. Variation in cardiac and demographic risk factors are likely to contribute to differences in the prevalence rates reported across different studies; however, our findings corroborate the observation that low levels of cognitive functioning are common in this patient group.
Previous studies have shown a relationship between cognitive functioning and inflammation. For example, Roberts and colleagues [
13] performed a case–control study in which inflammatory markers were compared in people with and without mild cognitive impairment. These authors reported that elevated levels of plasma CRP were significantly associated with mild cognitive impairment. In our study, we were able to assess the prospective association between cognition and inflammation and found that low cognitive function promotes increased inflammation in response to CABG surgery. There is a scarcity of research investigating inflammation and cognition in cardiac patients. A longitudinal study using participants taking part in the MacArthur Studies of Successful Ageing also reported a negative association between cognitive functioning and IL-6 over time, although this was a non-clinical sample and therefore not directly comparable to our findings [
15]. In our study, we only found evidence for low cognitive performance being associated with IL-6 and not CRP. The reason for this is not clear, although one possibility is that of timing. IL-6 is secreted by T cells and macrophages as part of the initial immune response and triggers the release of CRP from the liver. By measuring IL-6 and CRP at the same time, we may have captured these markers at different points in their decline towards baseline values. However, our findings warrant replication in further, large-scale studies of cardiac patients. Cognitive function explained 3.9 % of variance in IL-6 over and above other clinical, demographic and mood variables. We believe our findings to be of clinical relevance in that the heightened inflammation we observed in low cognitive function participants may in turn lead to phenomena that produce poorer recovery, such as slower wound healing or infection, or indeed, it could be important to later cardiovascular morbidity in the same way it is in acute coronary syndrome patients [
38,
39]. Greater inflammation after CABG has previously been associated with greater morbidity including increased risk of atrial fibrillation [
40], poorer lung function [
41] and even death [
42]. This hypothesis requires further testing using other longer-term CABG morbidity end points. We have shown in other analyses that the magnitude of acute inflammatory responses following surgery predicts later depressive symptoms [
43] and length of hospital stay [
44].
Previous research has shown that cognitive function, particularly memory, and cortisol levels are related [
45‐
47]. Whilst this literature is well established, few studies have investigated these effects in CABG patients. One study examined 166 patients after CABG surgery and found that higher levels of serum cortisol on the morning after surgery was associated with poorer cognitive performance 7 days later [
27]. Salivary cortisol has been found to increase greatly after non-cardiac surgery in association with post-operative cognitive dysfunction [
48], so this effect might not be specific to CABG surgery patients. In our study, we found that low pre-operative cognitive function predicts greater cortisol output across the day up to 2 months after CABG surgery. Indeed, cognitive function predicted 2.6 % of variance in cortisol AUC responses over and above covariates. There is limited evidence for the detrimental effects of elevated cortisol in this patient group; however, there is some evidence that it is associated with poorer lung function [
41] and increased risk of delirium [
49]. Moreover, we recently showed that flatter cortisol slopes prior to CABG were associated with greater mortality and major adverse events up to 2.68 years after surgery [
50]. We found an effect for cortisol AUC but not for cortisol awakening response or slope. The reason for this is not entirely clear, although previous work has suggested that the awakening response in particular is highly influenced by situational factors [
51], making it a less reliable indicator of trait responses. Further research is needed to understand whether these effects persist in the longer term and whether such elevations in cortisol translate into greater clinical risk in these patients.
Interestingly, we showed that not only those participants falling below the cutoff for mild cognitive impairment were at risk of poorer biological responses but also that cognition and inflammation and cortisol were associated in a graded fashion, indicative of a dose–response relationship. This suggests that even amongst participants within the normal range of cognitive performance, those scoring at the lower end of the scale had worse biological responses than those at the higher end of the scale. More work is needed to understand the use of cutoffs in this patient group and the benefits of early identification of patients with less than optimal cognitive functioning.
In the current study, heightened inflammatory and neuroendocrine activity after the stress of CABG surgery seen in patients with lower cognitive functioning suggest that there may be disruption of the HPA axis in some patients. The human stress response involves activation of the HPA axis, the subsequent release of cortisol and simultaneous elevation of inflammatory cytokines [
52]. However, we were unable to support the association between cortisol and inflammation using our data, perhaps due to the temporal differences in collection of blood and saliva samples. Further research examining these biomarkers in CABG patients is needed. In addition, non-biological mechanisms also warrant further examination since behavioural pathways including the role of physical activity, adherence to medications, diet and alcohol consumption may also be relevant. Psychological mechanisms are also likely to be important. It is possible that those with cognitive impairment are less likely to employ adaptive coping responses (e.g. social support) to a stressor such as CABG surgery, resulting in greater perceived stress and in turn poorer biological responses to surgery. These pathways require further study as potential mediators or moderators of this relationship.
There are several strengths to our study. The longitudinal design of the ARCS study allows for the temporal relationship between cognitive function and biological responses to surgery to be analysed. Moreover, the large range of psychosocial and clinical data collected from the ARCS participants has allowed us to control for multiple covariates that might confound the association between cognition and biomarkers. However, there are also a number of limitations. Firstly, we have relied on a short, global assessment of cognitive function that does not allow us to study the relationship between different cognitive domains and inflammation and cortisol. In addition, cognitive function was only assessed pre-operatively, so it was not possible to assess the effect of cognitive deterioration from before to after the surgical procedure. Moreover, missing data on the biological measures meant that we had to perform some of our analyses on a reduced sample size. The ARCS study had a relatively homogenous sample, being predominately male and of White ethnic origin making generalisation to other groups difficult. Only 9.4 % of our samples were female. This male majority is characteristic of the CABG surgical population more generally, with men more likely to receive a revascularisation procedure than women in the UK. However, we included sex in all analyses as a covariate in order to address this issue.
In conclusion, we found that low cognitive functioning prior to CABG surgery was associated with greater IL-6 responses in the days following surgery and greater cortisol output up to 2 months later. The poorer biological responses observed in these patients may help explain the link between cognitive decline and cardiovascular pathology.