Neuroimaging of frailty, delirium and postoperative cognitive dysfunction

Abstract

The objectives of this thesis were to assess the relation between brain MRI markers and frailty, to identify brain MRI markers that predispose to delirium and/or postoperative cognitive dysfunction (POCD) and to investigate whether postoperative delirium or POCD is related to preoperative to postoperative MRI brain changes. The first part of this thesis focused on frailty. The second part of this thesis focused on the association between MRI features and delirium and POCD. In chapter 4, we provided an overview of previous research on these topics. This overview was divided into research on MRI markers of neurodegenerative diseases, and MRI markers of neurovascular diseases. Overall, MRI markers of neurovascular diseases showed the most consistent association with delirium and POCD. In Chapter 5 we assessed preoperative brain MRI markers of neurodegenerative and neurovascular diseases in relation to postoperative delirium, in a large cohort of older patients that had major elective surgery, of whom 17% developed postoperative delirium. No association was found between preoperative brain volumes, white matter hyperintensities, cerebral perfusion and postoperative delirium. However, a non-significant trend towards an association between cortical brain infarcts and delirium, and a non-significant trend between shape of white matter hyperintensities and delirium was found. These findings suggest that patients with preoperative cortical brain infarcts and those with a more complex shape of white matter hyperintensity may be at an increased risk for developing delirium after major surgery. Chapter 6 shows a novel method, in which a hierarchical clustering approach was performed to analyze brain MRI markers of neurodegenerative and neurovascular diseases. The aim was to identify different MRI phenotypes of the brain in older patients scheduled for major elective surgery, and to assess the relation between these phenotypes and postoperative delirium. Six distinct MRI phenotypes of the brain were found: limited burden (three groups), mainly atrophy, mainly small vessel disease and atrophy, and multi-burden. Patients with a multi-burden phenotype showed a higher odds for developing postoperative delirium, suggesting that MRI phenotypes of the brain could assist in an improved understanding of the structural correlates of predisposition to postoperative delirium. In chapter 7 the relation between cerebral microbleeds and postoperative delirium and/or POCD was assessed in a group of 65 patients. No association was found between preoperative cerebral microbleeds and the occurrence of postoperative delirium and POCD. Furthermore, new postoperative cerebral microbleeds at three months follow-up were not associated with occurrence of postoperative delirium and POCD. Chapter 8 assessed the relation between occurrence of postoperative delirium and preoperative to postoperative brain changes. Patients showed a greater decrease in gray matter volume compared to control participants. In the patient group, patients with postoperative delirium showed a greater decrease in gray matter volume compared to patients without postoperative delirium, with an additional loss of around 5 ml in approximately 4 months. Furthermore, in the patient group, occurrence of postoperative delirium was associated with new postoperative brain infarcts. We concluded that a greater decrease in gray matter volume and occurrence of new brain infarcts in the postoperative period could be part of the underlying structural correlates of adverse cognitive outcomes after postoperative delirium. In conclusion, brain changes on MRI that predispose to postoperative adverse neurocognitive outcomes are probably multifactorial, resulting from both neurodegenerative and neurovascular diseases. Postoperative delirium is also related to new postoperative brain changes of neurodegenerative and neurovascular origin.