Mild cognitive impairment and kidney disease: clinical aspects

Chronic kidney disease (CKD) is now seen as a systemic disease involving also the central nervous system [1], but the link between the kidney and different organ systems and disease went unnoticed for a long time. The king of Poland, Stephen Bathory (1533–86), suffered from CKD due to polycystic kidney disease and depression [2]. Similarly, Wolfgang Amadeus Mozart was also thought to have had CKD [3] and depression [4]. A list of ‘Famous People Who Have Died from Kidney Disease’ [5] includes many who suffered from both CKD and depression or other signs of mental illness. Is this a coincidence or actually evidence of a link between kidney disease and brain dysfunction? This is not merely an academic question because all forms of mental illness can seriously impair an individual’s quality of life, and are frequently associated with progression of diseases and premature mortality, so it is worth the effort of trying to answer it. Europe and much of the industrialized countries are experiencing growing numbers of patients with CKD within their ageing populations [6]. CKD is complex and potentially fatal: (i) all organs are affected, sooner or later; (ii) the balance of plasma volume, electrolytes, acid–base and minerals, metabolites, hormones and proteins is disturbed; and (iii) patients often need a multidisciplinary team approach managing complex comorbidities, drug regimens and special diets. Although the prognosis of patients with CKD remains poor, their increasing life expectancy has shifted medical attention from life-threatening emergencies to long-term complications and sequelae, and how to improve quality of life [7]. Indeed, kidney failure has detrimental effects on health-related quality of life (HRQoL), reaching levels similar to those seen in patients with metastatic cancer [8]. This might be due to psychological factors, both kidney disease and cancer being chronic diseases with a bad prognosis. However, although the effect of CKD on quality of life is more evident in advanced stages (stage G4+) and in older patients [9, 10], a large study has shown a significant decrease in HRQoL as early as CKD stage G2 [11]. Notably, neurological and cognitive impairments [12], and depression [13] are among the most debilitating consequences of CKD contributing to the significantly reduced HRQoL [14]. The recognition of the association of uraemia with brain dysfunction goes back to the 1930s when Toulouse, Marchand and Courtois postulated ‘a specific disease entity with azotemia as one of the most characteristic symptoms: azotemic acute psychotic encephalitis’ [15]. The disease was subsequently renamed ‘uraemic encephalitis’ or ‘uraemic encephalopathy’ [16], including ‘the full spectrum of organic brain syndromes, progressing from mild impairment of intellect to coma’ [17]. At the same time, post-mortem studies in humans, particularly by Olsen [18], and in animal models were devoted to identifying the underlying brain abnormalities. The major anatomical changes observed in uraemia were white matter abnormalities, brain atrophy and neuronal degeneration [17]; functional alterations were also observed, consisting of changes in the permeability of the blood–brain barrier with resultant brain oedema [19], and altered content of neurotransmitters [20]. Initially, ‘uraemic encephalopathy’ was thought to be reversible by haemodialysis based on the assumption that it was caused by retained small molecule toxins. Already in 1967 Fishman and Raskin, however, predicted that ‘it is very likely that no single toxic compound will be identified’ [17], meaning that brain alterations might be due to multiple individual and additive causes. At the end of the 1960s, another neurological state, caused by aggressive dialysis in advanced uraemic states, was identified and called (dialysis) ‘disequilibrium syndrome’ [21]. In the 1970s, a peculiar form of dementia associated with chronic dialysis, called ‘dialysis dementia’, was also described and attributed to trace metals, most likely aluminium [22]. In the 1990s, significant improvements in dialysis methods and regimens were achieved, and the occurrence of advanced, untreated uraemic states was significantly reduced. The incidence of ‘disequilibrium syndrome’, ‘dialysis dementia’ and ‘uraemic encephalopathy’ was beginning to decline and attention shifted to the potentially beneficial brain effects of epoetin and growth hormone in CKD [23]; probably due to the wider availability of these new therapeutic options. Furthermore, as dialysis outcomes were steadily optimized, in part due to working groups such as DOQI in the 1980–90s, the meaning of the term ‘uraemia’ was changing from the terminal stages of body intoxication to the ‘residual syndrome’ defined by Depner [24] as the effects were not corrected by dialysis and possibly the result of other retained metabolic waste products. With the success of the KDOQI classification of CKD in 2002, the term uraemia was confined to the terminal condition of (often more acute) intoxication. Dialytic techniques then focused on the removal of medium-sized molecules, such as β-2 microglobulin, responsible for amyloid deposition in tissues, including the brain [25]. Unfortunately, no data are available yet on the prevention of mild cognitive impairment (MCI) using these new techniques. At the same time, milder neurological disease states, and more specifically cognitive decline in earlier CKD stages, came into focus. Thus, current studies no longer consider ‘uraemic encephalopathy’ as such, but are more commonly concerned with the recognition and nature of cognitive decline in CKD. Cognitive decline is one of the behavioural manifestations of brain damage in CKD. Other manifestations of brain dysfunction are sleep disturbances and depression, which represent a wide topic that is not covered in the present review. Patients with CKD show high prevalence of depression that can often appear as a reversible form of cognitive impairment, hence it is often called ‘pseudodementia’ or ‘functional dementia’ [26]. The identification of this form is difficult and the reader should refer to more specialized reviews (see e.g. [27]). Depression might itself be an effect of cognitive impairment. Prototypical forms of dementia such as Alzheimer’s disease and dementia with Lewy bodies, are, in fact, accompanied by a high prevalence of depression [28]. Furthermore, cognitive impairment usually persists after significant improvements in depressive symptoms (see e.g. [29]). Therefore, the coexistence of depression and MCI-CKD might be part of a more general phenomenon, rather the presence of two different nosological entities. Similarly, the association of peripheral neuropathy and cognitive impairment in patients with CKD can be observed in other conditions such as diabetes [30], vitamin B12 deficiency [31] and amyloidosis [32]. These associations wait to be better understood in the population of patients with CKD. The cognitive decline can manifest with a continuum from mild involvement, or MCI, up to clinically relevant dementia, when interference with daily life and independency is present. Being a prodromal stage, MCI should be identified and studied in CKD before irreversible damage is present. MCI is characterized by mild impairment in several key cognitive domains, that is, executive functions of memory (learning and attention), problem solving (processing) and self-control (emotion—depression). These domains can be captured by screening tests, for example, the Montreal Cognitive Assessment (MoCA), and neuropsychological tests, often leading to the diagnosis of MCI (Figure 1). The MoCA has been specifically designed for the detection of MCI, with a sensitivity of 80–100% and specificity of 50–76% using a cut-point of 25/26 [33]. Clinical testing of MCI in CKD patients. The diagnosis of MCI requires normal execution of routine tasks, and without interference with daily activity. Neuropsychological testing can be useful for screening although a full diagnosis requires further documentation. The neuropsychological tests usually evaluate multiple cognitive domains that are summarized in a single score. The MoCA has been validated as having better sensitivity then Mini Mental State Examination for MCI screening. Routine screening for MCI in CKD patients is not yet recommended, notwithstanding the high prevalence of the condition. This could be partially explained by the lack of intervention strategies once the condition has been recognized. However, the nephrologist should suspect cognitive impairment when the patient (or the caregiver) reports forgetfulness/confusion about medications, appointments or inappropriate calls, or when the patient reports depression or altered sleep patterns, and when the patient is not able to answer questions without the family member/caregiver, and in all cases of history of stroke or unexplained falls. MCI is a transitional stage, preceding clinical dementia, with cognitive impairment exceeding normal ageing and age-associated decline [34]. The main difference between dementia and MCI is that in the former the severity of cognitive impairment usually interferes with activities of daily living; however, it is almost certainly a spectrum with evidence that modifiable cardiovascular risk factors, physical exercise and cognitive training can slow progression [35]. In recent years, several clinical studies have concentrated on the burden of MCI in CKD patients (see e.g. [36–38]). Among patients with CKD, the prevalence of MCI has been estimated to be as high as 30% [39] to 63% [40], which is approximately twice as high as in the age-matched general population. The stage of CKD is apparently related to the risk of MCI: the lower the estimated glomerular filtration rate (eGFR) the higher the risk of MCI (see also Table 1). Notably, discernible cognitive changes may appear already in early CKD stages [51]. Prevalence of MCI and dementia in different populations Prevalence of MCI and dementia in different populations Interestingly, patients with end-stage renal disease (ESRD) have a similar MCI burden with or without haemodialysis, or peritoneal dialysis (Table 1). This is surprising because recent data show that haemodialysis modifies cerebral blood flow as a function of ultrafiltration volume [52]. However, after 12 months, this brain haemodynamic effect correlates with a better performance with the MoCA test for MCI [53]. This observation is also supported by the improvement of the cognitive functions after the start of dialysis, possibly with better results using peritoneal dialysis compared with haemodialysis [54]. Therefore, the decline of cerebral blood flow observed during haemodialysis may not contribute to MCI. This might be interpreted as the presence of toxins not eliminated by the dialysis process, or the deficit of (as yet unknown) neuroprotective substances produced by the kidney that have not been directly replaced yet. Much effort has been devoted to identifying specific risk factors for MCI in patients with CKD. To address this problem, investigators had to control several risk factors known for MCI in the general population, such as age, a family history for MCI, the education level, exposure to environmental chemicals, physical inactivity, diabetes, heart disease, stroke or other past brain injuries, male gender, hypertension, smoking, weight and hypercholesterolaemia [55]. Genetic susceptibility towards MCI, such as apoE genotype, is likely to play the same role in the general population and in CKD patients [56]. Nephrologists can easily recognize that many of these factors accompany CKD itself. However, when these confounders are taken into consideration, other CKD-specific risk factors can be identified such as albuminuria [57], normalized brain tissue volume, haemoglobin levels, glycaemia, serum parathyroid hormone and uric acid levels [58]. Another CKD-specific risk factor for MCI is the frequent presence of electrolyte disorders. Among these, hyponatraemia is relatively frequent in ESRD and has been associated with lower scores for tests of MCI [59]. However, it should be underlined that hyponatraemia is also a risk factor for MCI in non-CKD patients [60]. Other possible CKD-related factors for MCI are the duration of CKD, poor nutrition/protein energy wasting, neuro-psychological aspects, functional impairment, anaemia, acidosis, disturbed sleep and polypharmacy [61], although they are difficult to assess and have not been explored in any detail. The relative contribution of these and other factors and the still poorly understood underlying mechanisms remain major obstacles to prevention and treatment. In the population of CKD patients under dialysis, additional risk factors for MCI are related to the process and adequacy of renal replacement therapies. Specifically, risk factors for MCI are the total number of dialysis-related hypotension events [62] and, counterintuitively, a high equilibrated Kt/V >1.2 [39]. With respect to the dialysis modality, the risk for MCI is lower when the patients are placed on peritoneal dialysis or a central venous catheter is avoided [63]. However, the effects of different dialysis modalities on MCI could be spurious because of a selection bias of the patients, since randomization is usually not feasible. Kidney transplantation appears to reduce MCI (see Table 1), and this change is likely to be stable at 1 and 2 years after transplantation, suggesting the potential for some reversibility [64]. The reasons for this effect are still unclear. An attractive and untested hypothesis is that the kidney produces neurotrophic factors that are necessary for normal cognition in the long term. However, several other hypotheses are equally possible at this stage. Recent data suggest that a specific mechanism of brain impairment in CKD may be linked to chronic inflammation [13], altered levels of serum uric acid, parathyroid hormone levels, low glycaemia [58] and of neuropeptide Y (NPY), a sympathetic neurotransmitter thought to be involved in cognition and memory [65] and dysregulated in CKD [50]. The relevance of NPY is partially supported by the presence of autonomic dysfunction in CKD [66]. MCI itself is a risk factor for mortality, as already mentioned, and contributes significantly to the poorer quality of life of CKD patients [61]. MCI represents highly heterogeneous phenotypes with different underlying aetiologies (Figure 2). The MCI diagnosis relies (at least in part) on questionnaires testing multidimensional abilities of the patient (arithmetic, basic motor skills, time and place perception, repeating word list, language use and comprehension). This multidimensional testing procedure may lead to the grouping of different disease entities under a single heading of ‘MCI’. MCI subtypes. According to the number of cognitive domains involved, and the presence of memory impairment, MCI is further sub-classified, which can partially separate MCI into different aetiologies. Indeed, the concept of MCI was developed in the field of dementia, particularly for the population at risk of Alzheimer’s disease. Only much later, this concept has been applied to the study of cognitive changes in CKD. Therefore, it is very instructive to use the large body of neurophysiological and imaging data available from patients with MCI in the general population and to compare these with the MCI pattern found in the population of CKD patients. To distinguish these two populations, we will use the term MCI-GP when referring to the general population and MCI-CKD when referring to the CKD population. A large body of data comes from electroencephalographic (EEG) recordings. The first EEG studies in CKD were performed in the early 1960s [67, 68]. In a recent study comparing EEGs from MCI-GP with MCI-CKD, subtle differences emerged between the two populations: MCI-GP was characterized by alterations in the MCI-CKD was characterized by alterations in the The of these differences is the in the is present during and is thought to represent by the to the the is present during sleep and is thought to from between the and the Therefore, MCI-GP might be characterized by an altered in MCI-CKD an altered might although this is data based on imaging results the brain alterations in Unfortunately, no study directly compared patients with MCI-GP and MCI-CKD using data on the brain in CKD were available on post-mortem the first imaging study in patients with CKD, using was in the and focused on uraemic the of and reversibility with dialysis Recent data show a in matter in with of white matter in MCI-GP MCI-CKD with white matter particularly in white and reduced matter volume to the damage the of these differences is and toxic effects of uraemic substances be An advanced called more of in the white the of major brain as the and is also called between MCI-GP and MCI-CKD and so we on separate In in white matter were found in and and major with lower at the of the in MCI-CKD, in the of the and changes functional imaging data on brain levels or have been to brain in MCI-GP and MCI-CKD, but without The first study of the brain in uraemia was in on the data in cerebral blood flow in patients with MCI-CKD and those with this has not been observed in patients with on brain comes from animal animal models of MCI have been including or and at an early before brain These show of atrophy and damage to the system animal models of CKD or high with cognitive impairment show altered sleep with normal The imaging and animal data suggest differences between MCI-GP and MCI-CKD (see Table 2). These differences are likely to from the underlying of cognitive impairment in with and without CKD. MCI-GP is likely to be a clinical entity from MCI-CKD, the should be as a However, testing of this difference has not yet been and further questions are still is the of MCI-CKD, the same it is known to Is the CKD stage Is the of CKD relevant is evidence that albuminuria and are associated with functional and of MCI-CKD functional and of MCI-CKD an this from ‘uraemic because it appears before the uraemia and not improve with might be reversible although data on this are still it is a milder neurological to be a very early stage of may clinical In that it is a potential and in the CKD population and to be Although the first of uraemic was some years of brain dysfunction in CKD, prevention and is still in The MCI-CKD remains in the of hypothesis and specificity when compared with MCI-GP is based on the problem has increasing attention among as shown by a increasing number of neurological and psychological changes in CKD the number of the and was up to the in the years they are from The main in of this in the years has been more of a change in from uraemic to we and this a growing recognition of the problem of MCI, it also to full of earlier data and will to we are the main questions worth with the of the between MCI-GP and MCI-CKD or MCI-CKD represent an or of is the contribution of and cardiovascular risk factors to we have for the diagnosis of is the of MCI-CKD be or we assess the and burden of should we be more about an in this are at least two one is the and the is the the of MCI-CKD and MCI-GP is and is likely to of all forms of MCI. At the to and MCI is almost or physical exercise can be but are of dialysis techniques in patients with MCI-CKD represent an to evaluate these potential therapeutic Furthermore, most studies have been focused on the association of albuminuria and MCI, but it is that other of kidney function could be useful in the link the involving and inflammation in CKD further the new have available that may new on the early identification and of These techniques to and to their and brain have not been in this particularly to compare MCI-CKD and The of these techniques to and functional imaging of the brain in with neuro-psychological testing is a Furthermore, new animal models are now available that brain at the single in with proteins that are or that change their with to and neuronal populations, and also that can be using such as and should to some of the major of imaging techniques. The can an to the of in detail. systems for drug identification and neuronal to and test large of potential uraemic as as new and their The of from patients and brain could represent a new in for the and reversibility of MCI. The availability of a large of data that may or the occurrence of MCI and neurological new techniques to such as and and the of a new of These new and the to a new and better of MCI-CKD, as as MCI-CKD will an increasing problem by the and it is that we with clinical in and as as basic in to address this major and The that they have no of related to this