The journey through a disease such as dementia is one of the most complex that exists. Dementia’s character of global affectation and its irreversibility in the vast majority of cases produces an enormous impact not only on the patient, but on all those who must accompany the person on such a difficult trip. Among many other issues, we must face the lack of enough knowledge about its different etiologies, the complexities of diagnosis, and the absence of effective treatments. Thus, once the diagnosis arrives, all of this creates a “perfect storm” that affects patients and also their families, caregivers, and professionals who look after them.
Having crossed the threshold of a diagnosis that is not always simple and comes often later than desired, a question inevitably arises: What will be the evolution of this case? It is a question that understandably families and patients ask the doctor and that the clinician also asks himself. The issue of prognosis is always important in medicine, but in the case of dementia, its implications are very relevant given the progressive nature that the illness entails. These implications affect the patient’s quality of life in the years to come, their need for formal and informal care, unavoidable legal aspects, as well as the costs derived from their care. Will it progress quickly or slowly? What is the possibility of evolving with a large or limited number of symptoms compared to the wide spectrum of possible ones? How long will the person live? In short, what will be the trajectory of each patient, in particular, while navigating through storms that will never know a true calm?
Attempts to foresee the evolution of dementia are not new, but almost any approach to the published information brings with it a word: heterogeneity. Patients’ evolutions are very different in the speed of progression, severity, or the number and type of symptoms that occur. This heterogeneity is the result not only of the characteristics of the disease itself, but also of those of patients and their personal environment. Among many other factors, early versus late onset, time to diagnosis, different cognitive and behavioral profiles, genetic factors, health status, frailty, lifestyle, family and social support, and access to health and care services have been signaled.
Experience dictates that the approximations from the statistical mean are not very informative for any of the agents involved. On the other hand, the symptoms of dementia are various and may not evolve in parallel. Cognitive deterioration is the hallmark of dementia. However, the question of whether it is the most important factor in the different phases of the disease arises. In this line, some other questions such as the role of neuropsychiatric symptoms (NPS) – which may or not follow the same course of cognition – or the impact of present or later acquired physical comorbidity on evolution can be posed.
Although it is known that the factors that can influence evolution are multiple, researchers have generally focused on only one part of them, mostly cognition. In addition to pointing out heterogeneity, research results have at times provided contradictory data, as we will cite later.
Besides all this, measuring instruments are always a recurring problem in dementia research. Which ones are appropriate to capture the symptomatic evolution over time? Are they the same as for cross-sectional studies? Are measures as simple as the Mini-Mental State Examination (MMSE), on which many studies rely, a good correlate of patient’s evolution? Few investigators doubt that for now the measuring instruments, as a necessary compass that research studies on dementia need, were and still are insufficient, and therefore the journey has often taken deviant when not frankly wrong routes.
Finally, the methodology for analyzing the data obtained is also important. In this field, innovations have been experienced, with the incorporation of statistical techniques that go far beyond the analysis of multiple correlations. One of the most modern methodologies in the study of the long-term evolution of dementia is the one that uses the identification of latent classes and defines certain evolutionary trajectories based on the data obtained in the baseline of the study. We definitely need more math to combat the uncertainties of navigating these seas.
Knowing data about the evolution and prognosis of dementia is not only important for the management and planning for the future of patients. These data are also important for the design of clinical trials aimed at a utopian cure for the disease or at least at significantly slowing down its evolution. The heterogeneity that we outline here has undoubtedly been responsible for part of the limited progress obtained so far in the field of treatment. This has been widely recognized, for example, in the case of clinical trials conducted in patients with Mild Cognitive Impairment, in which the absence of sufficiently homogeneous cohorts has been usually the rule (Russ and Morling, Reference Russ and Morling2012).
On the other hand, there is a need to carry out clinical trials of a sufficient duration to detect significant changes in patients with established dementia, but the budget and attrition rate grows exponentially with time of follow-up. Knowing the characteristics of patients with a rapid progression could reduce this time, saving costs, and making this research simply feasible.
The work by Edwin et al. (Reference Edwin, Strand, Persson, Engedal, Selbaek and Knapskog2020) that we are commenting on here makes a 3-year follow-up of a sample of 668 patients seen in a memory clinic in Norway. The main variable of their study is the change in the Clinical Dementia Rating scale, as a compilation of the cognitive and functional state. Patients received an average of 3.5 visits during that period. The study is notable for the amount of patient information that it includes. Thus, sociodemographic data, comorbid diseases, blood sampling, and CSF and/or neuroimaging biomarkers were recorded. Cognitive status was measured through a battery of neurocognitive tests that included, among others, the MMSE, clock drawing test, trail making test A and parts of the Consortium to Establish a Registry for Alzheimer's Disease battery. NPS were monitored through the Neuropsychiatric Inventory. The study included not only Alzheimer’s disease, but also Parkinson’s, Lewy bodies, and frontotemporal dementias. Their sample consisted of relatively young individuals (mean 70.5 years) with moderate forms of dementia (mean MMSE 23.1) and little comorbidity. Interestingly, those patients who could not be included in the follow-up period were older, lived alone, had poorer cognitive functioning and more NPS, as well as worse physical condition and higher mortality than those that took part in the study. This accounts for the biases of most investigations of this type and the difficulties of including a part of patients with more serious characteristics, that just for these reasons, are less likely to meet the requirements of a research study.
One of the strengths of this study is the use of group-based trajectory modeling techniques that define different evolutionary trajectories in which patients can be grouped. After different comparisons, the authors found three main trajectories in which patients were classified: one of slow evolution and two of faster evolution. The definition of a majority group of relatively slow evolution, compared to others in which it is more rapid is a finding that seems to be repeated in other investigations that use similar methodologies (Baker et al., Reference Baker2017; Haaksma et al., Reference Haaksma, Calderon-Larranaga, Olde Rikkert, Melis and Leoutsakos2018a; Leoutsakos et al., Reference Leoutsakos2015; Tschanz et al., Reference Tschanz2011) and is consistent with the clinical experience of most clinicians. However, the proportion of patients with slow or rapid evolution is not always consistent in the different studies (Wang et al., Reference Wang2019).
The authors find that the most frequent trajectory (43.2% of the total) include patients with the best baseline functioning and is the one that has a slower evolution compared to the other two groups that evolved faster (34.9 and 21.9%, respectively). Belonging to the more rapidly evolving groups was associated with older age, worse baseline cognitive status, and a greater burden of NPS. The one with the fastest evolution of the three had a greater baseline deterioration and a mortality rate four times higher than the group with the best evolution. It is of interest to note that the two fastest decline rate groups did not differ from the slowest in terms of sex, etiology of dementia, or physical comorbidity.
The type of vessel in which the patient undertakes the journey through dementia and how he is equipped is very different from one person to another, and hence the heterogeneity of the evolutions. The different causes of this heterogeneity are given by differences related to the patient, such as the person’s age and the age of onset of symptoms, sociodemographic, genetic, and comorbidity factors, as well as those linked to the disease itself and its impact on the subject, such as cognitive or behavioral state, as well as the data provided by the biomarkers (Melis et al., Reference Melis, Haaksma and Muniz-Terrera2019; Song et al., Reference Song2018). But Edwin et al.’s (Reference Edwin, Strand, Persson, Engedal, Selbaek and Knapskog2020) research shows that certain predictability is possible, knowing in an extensive way the basal state of the patient at the beginning of the trip.
One of the controversial issues has been the age at the onset of the disease. There is data that supports that the beginning at a younger age would be linked to a faster evolution, but also the opposite has been found like in Edwin’s work (Barnes et al., Reference Barnes, Bartlett, Wolk, van der Flier and Frost2018; Jutkowitz et al., Reference Jutkowitz, MacLehose, Gaugler, Dowd, Kuntz and Kane2017; Wattmo and Wallin, Reference Wattmo and Wallin2017). In addition, the evolutionary moment in which the diagnosis is made is also important. There is evidence that, for a significant number of patients, dementia would evolve slowly at the beginning, it would accelerate in moderate stages of the disease until crossing a threshold in which, already advanced, the rapidity of decline would be lower again (Song et al., Reference Song2018). Age and time of diagnosis are important, because the possible therapeutic interventions and the accessibility of patients to them may be different in one case or another.
Another controversial element is the effect of educational level on the progression of dementia. Although this work does not find differences, it has been pointed out that people with higher educational levels would have a faster decline, which challenges older findings based on the theory of cognitive reserve (Sona et al., Reference Sona, Ellis and Ames2013).
An important point to remember is that non-memory cognitive symptoms (for example, executive dysfunction) and NPS are associated with a greater rate of deterioration (Lanctot et al., Reference Lanctot2017; Musicco et al., Reference Musicco2010). As the authors point out, the evolution of cognitive and behavioral symptoms appears to follow different trajectories and the impact of individual NPS is not clear yet (Haaksma et al., Reference Haaksma2018b).
The presence of physical comorbidity and the consequent frailty has been repeatedly indicated as a factor associated with a faster evolution (Haaksma et al., Reference Haaksma2017), although in the present paper, the authors found no differences in the baseline situation.
Finally, it is interesting to note that practically no published study includes patients from the time of diagnosis to the time of death. For this reason, it is difficult to discern whether the factors that influence the speed of evolution are the same or have the same impact in the short and very long term. Its knowledge would require very long-lasting studies, taking into account not only the baseline conditions, but also the new events that are occurring over time, with frequent periodic evaluations and statistical methods capable of adequately analyze these evolutionary data. This is a great challenge, which we are obliged to face in our research agenda.
The results of present and future research should lead to the development of more effective and predictive models, in line with those that are currently available in other areas of serious diseases such as oncology. Their results would offer a better navigation chart to patients and those who care for them for this hard sailing through seas that are still too uncertain and plagued by storms, as long as the terra firme of healing cannot yet be reached.
You have
Access
The journey through a disease such as dementia is one of the most complex that exists. Dementia’s character of global affectation and its irreversibility in the vast majority of cases produces an enormous impact not only on the patient, but on all those who must accompany the person on such a difficult trip. Among many other issues, we must face the lack of enough knowledge about its different etiologies, the complexities of diagnosis, and the absence of effective treatments. Thus, once the diagnosis arrives, all of this creates a “perfect storm” that affects patients and also their families, caregivers, and professionals who look after them.
Having crossed the threshold of a diagnosis that is not always simple and comes often later than desired, a question inevitably arises: What will be the evolution of this case? It is a question that understandably families and patients ask the doctor and that the clinician also asks himself. The issue of prognosis is always important in medicine, but in the case of dementia, its implications are very relevant given the progressive nature that the illness entails. These implications affect the patient’s quality of life in the years to come, their need for formal and informal care, unavoidable legal aspects, as well as the costs derived from their care. Will it progress quickly or slowly? What is the possibility of evolving with a large or limited number of symptoms compared to the wide spectrum of possible ones? How long will the person live? In short, what will be the trajectory of each patient, in particular, while navigating through storms that will never know a true calm?
Attempts to foresee the evolution of dementia are not new, but almost any approach to the published information brings with it a word: heterogeneity. Patients’ evolutions are very different in the speed of progression, severity, or the number and type of symptoms that occur. This heterogeneity is the result not only of the characteristics of the disease itself, but also of those of patients and their personal environment. Among many other factors, early versus late onset, time to diagnosis, different cognitive and behavioral profiles, genetic factors, health status, frailty, lifestyle, family and social support, and access to health and care services have been signaled.
Experience dictates that the approximations from the statistical mean are not very informative for any of the agents involved. On the other hand, the symptoms of dementia are various and may not evolve in parallel. Cognitive deterioration is the hallmark of dementia. However, the question of whether it is the most important factor in the different phases of the disease arises. In this line, some other questions such as the role of neuropsychiatric symptoms (NPS) – which may or not follow the same course of cognition – or the impact of present or later acquired physical comorbidity on evolution can be posed.
Although it is known that the factors that can influence evolution are multiple, researchers have generally focused on only one part of them, mostly cognition. In addition to pointing out heterogeneity, research results have at times provided contradictory data, as we will cite later.
Besides all this, measuring instruments are always a recurring problem in dementia research. Which ones are appropriate to capture the symptomatic evolution over time? Are they the same as for cross-sectional studies? Are measures as simple as the Mini-Mental State Examination (MMSE), on which many studies rely, a good correlate of patient’s evolution? Few investigators doubt that for now the measuring instruments, as a necessary compass that research studies on dementia need, were and still are insufficient, and therefore the journey has often taken deviant when not frankly wrong routes.
Finally, the methodology for analyzing the data obtained is also important. In this field, innovations have been experienced, with the incorporation of statistical techniques that go far beyond the analysis of multiple correlations. One of the most modern methodologies in the study of the long-term evolution of dementia is the one that uses the identification of latent classes and defines certain evolutionary trajectories based on the data obtained in the baseline of the study. We definitely need more math to combat the uncertainties of navigating these seas.
Knowing data about the evolution and prognosis of dementia is not only important for the management and planning for the future of patients. These data are also important for the design of clinical trials aimed at a utopian cure for the disease or at least at significantly slowing down its evolution. The heterogeneity that we outline here has undoubtedly been responsible for part of the limited progress obtained so far in the field of treatment. This has been widely recognized, for example, in the case of clinical trials conducted in patients with Mild Cognitive Impairment, in which the absence of sufficiently homogeneous cohorts has been usually the rule (Russ and Morling, Reference Russ and Morling2012).
On the other hand, there is a need to carry out clinical trials of a sufficient duration to detect significant changes in patients with established dementia, but the budget and attrition rate grows exponentially with time of follow-up. Knowing the characteristics of patients with a rapid progression could reduce this time, saving costs, and making this research simply feasible.
The work by Edwin et al. (Reference Edwin, Strand, Persson, Engedal, Selbaek and Knapskog2020) that we are commenting on here makes a 3-year follow-up of a sample of 668 patients seen in a memory clinic in Norway. The main variable of their study is the change in the Clinical Dementia Rating scale, as a compilation of the cognitive and functional state. Patients received an average of 3.5 visits during that period. The study is notable for the amount of patient information that it includes. Thus, sociodemographic data, comorbid diseases, blood sampling, and CSF and/or neuroimaging biomarkers were recorded. Cognitive status was measured through a battery of neurocognitive tests that included, among others, the MMSE, clock drawing test, trail making test A and parts of the Consortium to Establish a Registry for Alzheimer's Disease battery. NPS were monitored through the Neuropsychiatric Inventory. The study included not only Alzheimer’s disease, but also Parkinson’s, Lewy bodies, and frontotemporal dementias. Their sample consisted of relatively young individuals (mean 70.5 years) with moderate forms of dementia (mean MMSE 23.1) and little comorbidity. Interestingly, those patients who could not be included in the follow-up period were older, lived alone, had poorer cognitive functioning and more NPS, as well as worse physical condition and higher mortality than those that took part in the study. This accounts for the biases of most investigations of this type and the difficulties of including a part of patients with more serious characteristics, that just for these reasons, are less likely to meet the requirements of a research study.
One of the strengths of this study is the use of group-based trajectory modeling techniques that define different evolutionary trajectories in which patients can be grouped. After different comparisons, the authors found three main trajectories in which patients were classified: one of slow evolution and two of faster evolution. The definition of a majority group of relatively slow evolution, compared to others in which it is more rapid is a finding that seems to be repeated in other investigations that use similar methodologies (Baker et al., Reference Baker2017; Haaksma et al., Reference Haaksma, Calderon-Larranaga, Olde Rikkert, Melis and Leoutsakos2018a; Leoutsakos et al., Reference Leoutsakos2015; Tschanz et al., Reference Tschanz2011) and is consistent with the clinical experience of most clinicians. However, the proportion of patients with slow or rapid evolution is not always consistent in the different studies (Wang et al., Reference Wang2019).
The authors find that the most frequent trajectory (43.2% of the total) include patients with the best baseline functioning and is the one that has a slower evolution compared to the other two groups that evolved faster (34.9 and 21.9%, respectively). Belonging to the more rapidly evolving groups was associated with older age, worse baseline cognitive status, and a greater burden of NPS. The one with the fastest evolution of the three had a greater baseline deterioration and a mortality rate four times higher than the group with the best evolution. It is of interest to note that the two fastest decline rate groups did not differ from the slowest in terms of sex, etiology of dementia, or physical comorbidity.
The type of vessel in which the patient undertakes the journey through dementia and how he is equipped is very different from one person to another, and hence the heterogeneity of the evolutions. The different causes of this heterogeneity are given by differences related to the patient, such as the person’s age and the age of onset of symptoms, sociodemographic, genetic, and comorbidity factors, as well as those linked to the disease itself and its impact on the subject, such as cognitive or behavioral state, as well as the data provided by the biomarkers (Melis et al., Reference Melis, Haaksma and Muniz-Terrera2019; Song et al., Reference Song2018). But Edwin et al.’s (Reference Edwin, Strand, Persson, Engedal, Selbaek and Knapskog2020) research shows that certain predictability is possible, knowing in an extensive way the basal state of the patient at the beginning of the trip.
One of the controversial issues has been the age at the onset of the disease. There is data that supports that the beginning at a younger age would be linked to a faster evolution, but also the opposite has been found like in Edwin’s work (Barnes et al., Reference Barnes, Bartlett, Wolk, van der Flier and Frost2018; Jutkowitz et al., Reference Jutkowitz, MacLehose, Gaugler, Dowd, Kuntz and Kane2017; Wattmo and Wallin, Reference Wattmo and Wallin2017). In addition, the evolutionary moment in which the diagnosis is made is also important. There is evidence that, for a significant number of patients, dementia would evolve slowly at the beginning, it would accelerate in moderate stages of the disease until crossing a threshold in which, already advanced, the rapidity of decline would be lower again (Song et al., Reference Song2018). Age and time of diagnosis are important, because the possible therapeutic interventions and the accessibility of patients to them may be different in one case or another.
Another controversial element is the effect of educational level on the progression of dementia. Although this work does not find differences, it has been pointed out that people with higher educational levels would have a faster decline, which challenges older findings based on the theory of cognitive reserve (Sona et al., Reference Sona, Ellis and Ames2013).
An important point to remember is that non-memory cognitive symptoms (for example, executive dysfunction) and NPS are associated with a greater rate of deterioration (Lanctot et al., Reference Lanctot2017; Musicco et al., Reference Musicco2010). As the authors point out, the evolution of cognitive and behavioral symptoms appears to follow different trajectories and the impact of individual NPS is not clear yet (Haaksma et al., Reference Haaksma2018b).
The presence of physical comorbidity and the consequent frailty has been repeatedly indicated as a factor associated with a faster evolution (Haaksma et al., Reference Haaksma2017), although in the present paper, the authors found no differences in the baseline situation.
Finally, it is interesting to note that practically no published study includes patients from the time of diagnosis to the time of death. For this reason, it is difficult to discern whether the factors that influence the speed of evolution are the same or have the same impact in the short and very long term. Its knowledge would require very long-lasting studies, taking into account not only the baseline conditions, but also the new events that are occurring over time, with frequent periodic evaluations and statistical methods capable of adequately analyze these evolutionary data. This is a great challenge, which we are obliged to face in our research agenda.
The results of present and future research should lead to the development of more effective and predictive models, in line with those that are currently available in other areas of serious diseases such as oncology. Their results would offer a better navigation chart to patients and those who care for them for this hard sailing through seas that are still too uncertain and plagued by storms, as long as the terra firme of healing cannot yet be reached.