As readers of this journal will be well aware, the last decade has witnessed a transformation in our understanding of the pathophysiology of dementias, perhaps most especially Alzheimer's disease (AD), as well as a much greater understanding of the undoubtedly complex and multifactorial etiology of the AD process. An increasing number of genetic risk and protective factors have been identified, as well as potentially modifiable risk factors, including vascular factors such as hypertension and exercise, education, lifestyle, and nutrition. In parallel, there has been considerable progress in developing and validating biomarkers for AD and other dementias. These have transformed the landscape for research, allowing in vivo patient stratification according to pathology, and now have fed through to inform our clinical diagnostic criteria. Particular examples include imaging biomarkers such as characteristic atrophy patterns on magnetic resonance imaging (MRI), hypoperfusion/hypometabolism on functional brain imaging, increased cortical amyloid uptake on positron emission tomography (PET), as well as CSF alterations in tau and β-amyloid. Several changes in blood have also been shown, including altered inflammatory markers, which may prove to be important biomarkers in the future.
However, several challenges remain. There is a need to understand how these different risk and protective factors and biomarkers relate to each other, and to cognitive decline. We also need to better understand the temporal sequencing of different aspects of the disease process, and determine the changes that represent core aspects of the pathophysiological process and are incidental bystanders. Perhaps most importantly, biological targets or risk factors that provide credible prospects for intervention should be identified, and greater knowledge of how such factors change over time at different points during the disease process will help clarify whether and how putative interventions might modify the underlying disease.
These are extremely challenging questions for the dementia field to answer. It has responded in a number of ways, including the formation of a several sizeable clinical cohorts with multiple assessments of clinical, cognitive, and biomarker characteristics followed over time. Such “deep phenotyping” of cohorts that are sufficiently large to enable complex interactions to be assessed, and contain sufficient participants who will cognitively progress to determine impact on disease process, are essential. The impact of such cohorts is exemplified by one of the earliest cohorts, that is the US Alzheimer's Disease Neuroimaging Initiative (ADNI) cohort, whose multiple publications are widely known and whose data have been made available and widely used by researchers around the globe.
In this month's issue of International Psychogeriatrics, Ellis and colleagues (Reference Ellis2014) publish a paper from the highly regarded Australian Imaging, Biomarkers and Lifestyle Flagship Study of Ageing (AIBL). This is indeed a “flagship” study, not just in name but in nature as well. The AIBL study consists of 1,112 patients comprising those with established AD, with mild cognitive impairment (MCI), and healthy older controls. Of particular note, the AIBL cohort predominantly focuses on the early transition from normality to cognitive impairment, as around three quarters of the participants in the cohort are healthy controls. It also provides a combination of a comprehensive and longitudinal clinical and biomarker assessment with a focus on prior lifestyle factors such as diet and exercise. Some findings from the AIBL study have already been published (Ellis et al., Reference Ellis2009; Rowe et al., Reference Rowe2010; Sona et al., Reference Sona2012). The paper by Ellis et al (2014) describes the 18-month outcome of participants in terms of diagnostic transition and cognitive change. There are several extremely interesting and important observations from this study, which have relevance for all cohort studies. The degree of ascertainment of follow-up was impressive, with nearly 90% returning for assessment at 18 months. Interestingly, of those who did not return, very few declined follow-up because of the intensity of the baseline assessment, suggesting that participants in such studies, certainly in the AIBL study, are deeply committed and highly motivated to participate in research. Similarly, mortality rate was low (around 0.3%), suggesting that these are a slightly fitter and more robust group than the general population. Of the most interest were the rates of cognitive decline or “transition” in diagnostic category. Of the healthy controls, 2.5% were re-diagnosed as MCI at follow-up, while 30.5% of those with MCI had progressed to dementia, all but one patient being diagnosed with AD. Diagnostic stability of AD and MCI was high, with only a small proportion (around 5%) of those with MCI “reverting” to normality. These findings are extremely important not only because it is clearly essential in such studies to have such diagnoses that are reliable but also the whole reason such cohorts are established is to determine factors associated with decline, so it is important that decline does indeed occur, and at a reasonably high rate. The 30% transition rate from MCI to AD over 18 months represents around a 20% per year rate, slightly higher but not out of keeping with the 12–15% per year reported from other studies. The authors are most likely right in their explanation of both the stability of diagnoses and the relatively high rates of decline from MCI to AD, in that recruitment was not epidemiological but from carefully selected excellent clinical services in both Melbourne and Perth where rigor of diagnostic decision-making is expected to be high. In addition, the AIBL study's panel method of consistent consensus clinical diagnosis for all the study participants is to be commended and, short of autopsy, represents the “gold standard” in the field for such studies.
The authors divided their healthy older controls into two groups (interestingly around half the participants were in each group), i.e. those with and without subjective memory complaints. As many previous studies have found, subjective memory complaints and their resolution over time were significantly associated with the presence of minor symptoms of depression and anxiety. Although numbers are small, it is also interesting to note that more patients in the subjective memory complaint group than in the non-complaint group progressed to MCI over the follow-up period. Finally, the paper provides information on cognitive factors associated with progression and, consistent with other literature, found that the MCI patients who progress to AD were slightly more impaired at baseline in cognitive measures, especially episodic memory and fluency.
So where do things go from here? Clearly, the combination of clinical, cognitive, risk factor, and biomarker data from the AIBL cohort will help provide some answers to several current questions about the key factors associated with cognitive decline and the rate and temporal ordering of clinical and biomarker changes. Challenges remain, even cohorts as sizeable as the AIBL study are relatively limited in size (partly because of the great expense of undertaking such longitudinal deep phenotyping studies) to fully address the investigation of all interactions between environmental, genetic, and lifestyle risk factors. Increased collaborative efforts both nationally and internationally are emerging and will be essential to make future progress. These will undoubtedly lead to the establishment of new cohorts, but progress may also be made by smarter use of existing data, including allowing open access and combining data, as well as fully harnessing the considerable potential that electronic health record systems, now introduced in several countries, give to allow in anonymized form greater research use of data captured for clinical purposes.
Conflict of interest
None.
You have
Access
As readers of this journal will be well aware, the last decade has witnessed a transformation in our understanding of the pathophysiology of dementias, perhaps most especially Alzheimer's disease (AD), as well as a much greater understanding of the undoubtedly complex and multifactorial etiology of the AD process. An increasing number of genetic risk and protective factors have been identified, as well as potentially modifiable risk factors, including vascular factors such as hypertension and exercise, education, lifestyle, and nutrition. In parallel, there has been considerable progress in developing and validating biomarkers for AD and other dementias. These have transformed the landscape for research, allowing in vivo patient stratification according to pathology, and now have fed through to inform our clinical diagnostic criteria. Particular examples include imaging biomarkers such as characteristic atrophy patterns on magnetic resonance imaging (MRI), hypoperfusion/hypometabolism on functional brain imaging, increased cortical amyloid uptake on positron emission tomography (PET), as well as CSF alterations in tau and β-amyloid. Several changes in blood have also been shown, including altered inflammatory markers, which may prove to be important biomarkers in the future.
However, several challenges remain. There is a need to understand how these different risk and protective factors and biomarkers relate to each other, and to cognitive decline. We also need to better understand the temporal sequencing of different aspects of the disease process, and determine the changes that represent core aspects of the pathophysiological process and are incidental bystanders. Perhaps most importantly, biological targets or risk factors that provide credible prospects for intervention should be identified, and greater knowledge of how such factors change over time at different points during the disease process will help clarify whether and how putative interventions might modify the underlying disease.
These are extremely challenging questions for the dementia field to answer. It has responded in a number of ways, including the formation of a several sizeable clinical cohorts with multiple assessments of clinical, cognitive, and biomarker characteristics followed over time. Such “deep phenotyping” of cohorts that are sufficiently large to enable complex interactions to be assessed, and contain sufficient participants who will cognitively progress to determine impact on disease process, are essential. The impact of such cohorts is exemplified by one of the earliest cohorts, that is the US Alzheimer's Disease Neuroimaging Initiative (ADNI) cohort, whose multiple publications are widely known and whose data have been made available and widely used by researchers around the globe.
In this month's issue of International Psychogeriatrics, Ellis and colleagues (Reference Ellis2014) publish a paper from the highly regarded Australian Imaging, Biomarkers and Lifestyle Flagship Study of Ageing (AIBL). This is indeed a “flagship” study, not just in name but in nature as well. The AIBL study consists of 1,112 patients comprising those with established AD, with mild cognitive impairment (MCI), and healthy older controls. Of particular note, the AIBL cohort predominantly focuses on the early transition from normality to cognitive impairment, as around three quarters of the participants in the cohort are healthy controls. It also provides a combination of a comprehensive and longitudinal clinical and biomarker assessment with a focus on prior lifestyle factors such as diet and exercise. Some findings from the AIBL study have already been published (Ellis et al., Reference Ellis2009; Rowe et al., Reference Rowe2010; Sona et al., Reference Sona2012). The paper by Ellis et al (2014) describes the 18-month outcome of participants in terms of diagnostic transition and cognitive change. There are several extremely interesting and important observations from this study, which have relevance for all cohort studies. The degree of ascertainment of follow-up was impressive, with nearly 90% returning for assessment at 18 months. Interestingly, of those who did not return, very few declined follow-up because of the intensity of the baseline assessment, suggesting that participants in such studies, certainly in the AIBL study, are deeply committed and highly motivated to participate in research. Similarly, mortality rate was low (around 0.3%), suggesting that these are a slightly fitter and more robust group than the general population. Of the most interest were the rates of cognitive decline or “transition” in diagnostic category. Of the healthy controls, 2.5% were re-diagnosed as MCI at follow-up, while 30.5% of those with MCI had progressed to dementia, all but one patient being diagnosed with AD. Diagnostic stability of AD and MCI was high, with only a small proportion (around 5%) of those with MCI “reverting” to normality. These findings are extremely important not only because it is clearly essential in such studies to have such diagnoses that are reliable but also the whole reason such cohorts are established is to determine factors associated with decline, so it is important that decline does indeed occur, and at a reasonably high rate. The 30% transition rate from MCI to AD over 18 months represents around a 20% per year rate, slightly higher but not out of keeping with the 12–15% per year reported from other studies. The authors are most likely right in their explanation of both the stability of diagnoses and the relatively high rates of decline from MCI to AD, in that recruitment was not epidemiological but from carefully selected excellent clinical services in both Melbourne and Perth where rigor of diagnostic decision-making is expected to be high. In addition, the AIBL study's panel method of consistent consensus clinical diagnosis for all the study participants is to be commended and, short of autopsy, represents the “gold standard” in the field for such studies.
The authors divided their healthy older controls into two groups (interestingly around half the participants were in each group), i.e. those with and without subjective memory complaints. As many previous studies have found, subjective memory complaints and their resolution over time were significantly associated with the presence of minor symptoms of depression and anxiety. Although numbers are small, it is also interesting to note that more patients in the subjective memory complaint group than in the non-complaint group progressed to MCI over the follow-up period. Finally, the paper provides information on cognitive factors associated with progression and, consistent with other literature, found that the MCI patients who progress to AD were slightly more impaired at baseline in cognitive measures, especially episodic memory and fluency.
So where do things go from here? Clearly, the combination of clinical, cognitive, risk factor, and biomarker data from the AIBL cohort will help provide some answers to several current questions about the key factors associated with cognitive decline and the rate and temporal ordering of clinical and biomarker changes. Challenges remain, even cohorts as sizeable as the AIBL study are relatively limited in size (partly because of the great expense of undertaking such longitudinal deep phenotyping studies) to fully address the investigation of all interactions between environmental, genetic, and lifestyle risk factors. Increased collaborative efforts both nationally and internationally are emerging and will be essential to make future progress. These will undoubtedly lead to the establishment of new cohorts, but progress may also be made by smarter use of existing data, including allowing open access and combining data, as well as fully harnessing the considerable potential that electronic health record systems, now introduced in several countries, give to allow in anonymized form greater research use of data captured for clinical purposes.
Conflict of interest
None.