Treatment Effects In Clinical Trials Research Articles

Abstract 4136785: Interpreting Population Mean Treatment Effects in the Kansas City Cardiomyopathy Questionnaire

Introduction: The Kansas City Cardiomyopathy Questionnaire (KCCQ) is a common outcome in heart failure trials. While comparing means between treatment groups improves statistical power, mean effects do not necessarily reflect the clinical benefit experienced by patients. Objectives: To evaluate the relationship between mean KCCQ treatment effects in clinical trials and the proportions of patients experiencing clinically important improvements across a range of heart failure etiologies and interventions. Methods: We performed a patient-level meta-analysis of 11 randomized trials (9,977 patients) to examine the relationship between mean KCCQ Overall Summary Score (OSS) treatment effects and the absolute differences in the proportions of patients experiencing clinically important (≥5 points) and moderate or larger (≥10 points) improvements. Validation was performed in 7 additional trials. Results: Group mean KCCQ-OSS differences were strongly correlated with absolute differences in clinically important changes (Spearman correlations 0.76-0.92). For example, a mean KCCQ-OSS treatment effect of 2.5 points (half of a minimally important difference for an individual patient) was associated with an absolute difference of 6.0% (95% prediction interval [PI] 4.0-8.1) in the proportion of patients improving ≥5 points and 5.0% (95% PI 3.1-7.0) in the proportion improving ≥10 points, corresponding to NNTs of 17 (95% PI 12-25) and 20 (95% PI 14-33), respectively. Similar relationships were observed in the validation studies. Conclusions: Inferences on clinical impact based on population-level mean treatment effects are misleading since even small between group differences may reflect clinically important treatment benefits at the patient level. Clinical trials should explicitly describe the distributions of patient-level KCCQ changes to support clinical interpretation of their results.

Heterogeneity of Treatment Effects in Clinical Trials: Overview of Multivariable Approaches and Practical Recommendations

Heterogeneity of Treatment Effects in Clinical Trials: Overview of Multivariable Approaches and Practical Recommendations

Longitudinal network-based brain grey matter MRI measures are clinically relevant and sensitive to treatment effects in multiple sclerosis.

In multiple sclerosis clinical trials, MRI outcome measures are typically extracted at a whole-brain level, but pathology is not homogeneous across the brain and so whole-brain measures may overlook regional treatment effects. Data-driven methods, such as independent component analysis, have shown promise in identifying regional disease effects but can only be computed at a group level and cannot be applied prospectively. The aim of this work was to develop a technique to extract longitudinal independent component analysis network-based measures of co-varying grey matter volumes, derived from T1-weighted volumetric MRI, in individual study participants, and assess their association with disability progression and treatment effects in clinical trials. We used longitudinal MRI and clinical data from 5089 participants (22 045 visits) with multiple sclerosis from eight clinical trials. We included people with relapsing-remitting, primary and secondary progressive multiple sclerosis. We used data from five negative clinical trials (2764 participants, 13 222 visits) to extract the independent component analysis-based measures. We then trained and cross-validated a least absolute shrinkage and selection operator regression model (which can be applied prospectively to previously unseen data) to predict the independent component analysis measures from the same regional MRI volume measures and applied it to data from three positive clinical trials (2325 participants, 8823 visits). We used nested mixed-effect models to determine how networks differ across multiple sclerosis phenotypes are associated with disability progression and to test sensitivity to treatment effects. We found 17 consistent patterns of co-varying regional volumes. In the training cohort, volume loss was faster in four networks in people with secondary progressive compared with relapsing-remitting multiple sclerosis and three networks with primary progressive multiple sclerosis. Volume changes were faster in secondary compared with primary progressive multiple sclerosis in four networks. In the combined positive trials cohort, eight independent component analysis networks and whole-brain grey matter volume measures showed treatment effects, and the magnitude of treatment-placebo differences in the network-based measures was consistently greater than with whole-brain grey matter volume measures. Longitudinal network-based analysis of grey matter volume changes is feasible using clinical trial data, showing differences cross-sectionally and longitudinally between multiple sclerosis phenotypes, associated with disability progression, and treatment effects. Future work is required to understand the pathological mechanisms underlying these regional changes.

Group sequential methods based on supremum logrank statistics underproportional and nonproportionalhazards.

Despite the widespread use of Cox regression for modeling treatment effects in clinical trials, in immunotherapy oncology trials and other settings therapeutic benefits are not immediately realized thereby violating the proportional hazards assumption. Weighted logrank tests and the so-called Maxcombo test involving the combination of multiple logrank test statistics have been advocated to increase power for detecting effects in these and other settings where hazards are nonproportional. We describe a testing framework based on supremum logrank statistics created by successively analyzing and excluding early events, or obtained using a moving time window. We then describe how such tests can be conducted in a group sequential trial with interim analyses conducted for potential early stopping of benefit. The crossing boundaries for the interim test statistics are determined using an easy-to-implement Monte Carlo algorithm. Numerical studies illustrate the good frequency properties of the proposed group sequential methods.

Scaling and interpreting treatment effects in clinical trials using restricted mean survival time.

Restricted mean survival time is the expected duration of survival up to a chosen time of restriction . For comparison studies, the difference in restricted mean survival times between two groups provides a summary measure of the treatment effect that is free of assumptions regarding the relative shape of the two survival curves, such as proportional hazards. However, it can be difficult to judge the magnitude of the effect from a comparison of restricted means due to the truncation of observation at time . In this article, we describe additional ways of expressing the treatment effect based on restricted means that can be helpful in this regard. These include the ratio of restricted means, the ratio of life-years (or time) lost, and the average integrated difference between the survival curves, equal to the difference in restricted means divided by These alternative metrics are straightforward to calculate and provide a means for scaling the effect size as an aid to interpretation. Examples from two randomized, multicenter clinical trials in prostate cancer, NRG/RTOG 0521 and NRG/RTOG 0534, with primary endpoints of overall survival and biochemical/radiological progression-free survival, respectively, are presented to illustrate the ideas. The four effect measures (restricted mean survival time difference, restricted mean survival time ratio, time lost ratio, and average survival rate difference) were 0.45 years, 1.05, 0.81, and 0.038 for RTOG 0521 and 1.36 years, 1.17, 0.56, and 0.12 for RTOG 0534 with = 12 and 11 years, respectively. Thus, for example, the 0.45-year difference in the first trial translates into a 19% reduction in time lost and a 3.8% average absolute difference between the survival curves over the 12-year horizon, a modest effect size, whereas the 1.36-year difference in the second trial corresponds to a 44% reduction in time lost and a 12% absolute survival difference, a rather large effect. In addition to the difference in restricted mean survival times, these alternative measures can be helpful in determining whether the magnitude of the treatment effect is clinically meaningful.

Protein biomarker signature in patients with spinal and bulbar muscular atrophy.

Spinal and bulbar muscular atrophy (SBMA) is a slowly progressing disease with limited sensitive biomarkers that support clinical research. We analyzed plasma and serum samples from patients with SBMA and matched healthy controls in multiple cohorts, identifying 40 highly reproducible SBMA-associated proteins out of nearly 3,000 measured. These proteins were robustly enriched in gene sets of skeletal muscle expression and processes related to mitochondria and calcium signaling. Many proteins outperformed currently used clinical laboratory tests (e.g., creatine kinase [CK]) in distinguishing patients from controls and in their correlations with clinical and functional traits in patients. Two of the 40 proteins, Ectodysplasin A2 receptor (EDA2R) and Repulsive guidance molecule A (RGMA), were found to be associated with decreased survival and body weight in a mouse model of SBMA. In summary, we identified what we believe to be a robust and novel set of fluid protein biomarkers in SBMA that are linked with relevant disease features in patients and in a mouse model of disease. Changes in these SBMA-associated proteins could be used as an early predictor of treatment effects in clinical trials.

Item-level heterogeneous treatment effects of selective serotonin reuptake inhibitors (SSRIs) on depression: implications for inference, generalizability, and identification

Abstract Objectives In analysis of randomized controlled trials (RCTs) with patient-reported outcome measures (PROMs), Item Response Theory (IRT) models that allow for heterogeneity in the treatment effect at the item level merit consideration. These models for “item-level heterogeneous treatment effects” (IL-HTE) can provide more accurate statistical inference, allow researchers to better generalize their results, and resolve critical identification problems in the estimation of interaction effects. In this study, we extend the IL-HTE model to polytomous data and apply the model to determine how the effect of selective serotonin reuptake inhibitors (SSRIs) on depression varies across the items on a depression rating scale. Methods We first conduct a Monte Carlo simulation study to assess the performance of the polytomous IL-HTE model under a range of conditions. We then apply the IL-HTE model to item-level data from 24 RCTs measuring the effect of SSRIs on depression using the 17-item Hamilton Depression Rating Scale (HDRS-17) and estimate heterogeneity by subscale (HDRS-6). Results Our simulation results show that ignoring IL-HTE can yield standard errors that are as much as 50 % too small and create significant bias in treatment by covariate interaction effects when item-specific treatment effects are correlated with item location, and that the application of the IL-HTE model resolves these issues. Our empirical application shows that while the average effect of SSRIs on depression is beneficial (i.e., negative) and statistically significant, there is substantial IL-HTE, with estimates of the standard deviation of item-level effects nearly as large as the average effect. We show that this substantial IL-HTE is driven primarily by systematically larger effects on the HDRS-6 subscale items. Conclusions The IL-HTE model has the potential to provide new insights for the inference, generalizability, and identification of treatment effects in clinical trials using PROMs.

Confidence distributions for treatment effects in clinical trials: Posteriors without priors.

An attractive feature of using a Bayesian analysis for a clinical trial is that knowledge and uncertainty about the treatment effect is summarized in a posterior probability distribution. Researchers often find probability statements about treatment effects highly intuitive and the fact that this is not accommodated in frequentist inference is a disadvantage. At the same time, the requirement to specify a prior distribution in order to obtain a posterior distribution is sometimes an artificial process that may introduce subjectivity or complexity into the analysis. This paper considers a compromise involving confidence distributions, which are probability distributions that summarize uncertainty about the treatment effect without the need for a prior distribution and in a way that is fully compatible with frequentist inference. The concept of a confidence distribution provides a posterior-like probability distribution that is distinct from, but exists in tandem with, the relative frequency interpretation of probability used in frequentist inference. Although they have been discussed for decades, confidence distributions are not well known among clinical trial statisticians and the goal of this paper is to discuss their use in analyzing treatment effects from randomized trials. As well as providing an introduction to confidence distributions, some illustrative examples relevant to clinical trials are presented, along with various case studies based on real clinical trials. It is recommended that trial statisticians consider presenting confidence distributions for treatment effects when reporting analyses of clinical trials.

Evaluating a shrinkage estimator for the treatment effect in clinical trials.

The main objective of most clinical trials is to estimate the effect of some treatment compared to a control condition. We define the signal-to-noise ratio (SNR) as the ratio of the true treatment effect to the SE of its estimate. In a previous publication in this journal, we estimated the distribution of the SNR among the clinical trials in the Cochrane Database of Systematic Reviews (CDSR). We found that the SNR is often low, which implies that the power against the true effect is also low in many trials. Here we use the fact that the CDSR is a collection of meta-analyses to quantitatively assess the consequences. Among trials that have reached statistical significance we find considerable overoptimism of the usual unbiased estimator and under-coverage of the associated confidence interval. Previously, we have proposed a novel shrinkage estimator to address this "winner's curse." We compare the performance of our shrinkage estimator to the usual unbiased estimator in terms of the root mean squared error, the coverage and the bias of the magnitude. We find superior performance of the shrinkage estimator both conditionally and unconditionally on statistical significance.

Routes to standardisation for tau protein measurement in CSF

AbstractBackgroundCSF biomarkers are the target of several commercially available in vitro diagnostic (IVD) kits allowing an early diagnosis of Alzheimer’s disease (AD) and eventually the monitoring of treatment effects in clinical trials. Tau hyper‐phosphorylation is one of the main hallmarks of AD and its measurement provides higher accuracy to stratify patients at an early‐stage.Due to the rapid advancements in proteomics, standardisation of measurements can be achieved through the development of reference measurement procedures (RMPs) and certified reference materials (CRMs).MethodSI‐traceable protein and peptide primary calibrators were developed for t‐tau and p‐tau: mass fraction was determined by amino acid analysis and purity was assessed by LC‐MS at high resolution. CSF pools at different tau concentration were used for the development and validation of LC‐IDMS candidate RMP and their commutability was assessed.ResultAn LC‐IDMS candidate reference method was develop and validated in CSF for t‐tau using a SI‐traceable candidate protein primary calibrator. Three different CSF pools (low, medium, high) were successfully quantified with a relative expanded uncertainty below 10%. The commutability of 13 matrix‐based CRMs was assessed involving 8 immunoassays from 5 IVD providers by measuring 40 single CSF donations.According to the AT(N) classification, high levels of p‐tau(181) are highly correlated to tau pathology and commercially available immunoassays are able to target this specific epitope. We conducted a feasibility study on the development of a candidate reference method for p‐tau(181) in CSF relying on a SI‐traceable peptide primary calibrator and an antibody‐free sample preparation workflow.New assays targeting p‐tau(217) in CSF and plasma were also developed, making it necessary the organisation of standardisation initiatives. We sourced and characterised multiple phosphopeptide primary calibrators as well as a GSK3beta‐phosphorylated protein primary calibrator for the development of an LC‐MS RMP including an immunoprecipitation step in the sample preparation workflow.Conclusiontau is considered as a priority for the IFCC WG on NDD and the development of SI traceable RMPs and CRMs is essential to underpin a worldwide standardisation of results. This will help IVD providers to address the traceability requirements of the IVDR thus facilitating regulatory approvals of the kits.

Patient-Reported Outcome Measures for Severe Recurrent Bilateral Nasal Polyps: Psychometric Evaluation and Content Validity.

To date, no patient-reported outcome measures have been specifically developed to assess pharmacological treatment effect in participants with severe chronic rhinosinusitis (CRS) with recurrent bilateral nasal polyps (NP). These studies aimed to assess (1) the psychometric properties and (2) content validity of Visual Analogue Scales (VAS) assessing NP symptom severity. (1) Retrospective psychometric validation study using clinical trial data and (2) cross-sectional qualitative patient interview study. (1) Multicentre trial; (2) real-world. (1) Psychometric validation was performed using data from a randomized, double-blind, placebo-controlled, Phase II study (NCT01362244) investigating the effect of mepolizumab in 105 participants with severe, recurrent bilateral NP currently needing polypectomy surgery. (2) Content validity was explored through cognitive debriefing interviews in 27 adults with severe CRS with recurrent bilateral NP who had received NP surgery in the past 10 years (NCT03221192). (1) Acceptable reliability, validity, and responsiveness were shown for individual VAS items, although the loss of smell VAS item performed poorly in several analyses, suggesting further evaluation of this item is needed. (2) All individual VAS items were well understood, considered relevant and were consistently interpreted by most participants, providing evidence for their content validity. These findings support the use of symptom VAS measures to evaluate disease experience and treatment effect in clinical trials of participants with severe CRS with recurrent bilateral NP.

Adherence, per-protocol effects, and the estimands framework.

In the statistical literature, treatment effects in clinical trials are frequently described as either ITT or per-protocol effects. The estimand given for the per-protocol effect is the effect in adherers, where adherers are typically defined as adhering to the intervention as specified in the trial protocol. This dichotomy of treatment effects is unhelpful when there are in reality multiple treatment effects that can be of clinical interest and relevance. The terms "per-protocol" and "adherence" are confusing to non-statisticians. Protocols always allow for discontinuation of randomized treatment so participants discontinuing have actually followed the protocol. When rescue or additional medication is available, the effect in adherers could mean the effect regardless of use of these medications or the effect in a counterfactual world where the participant did not take the medication. Adherence can mean continuing to be prescribed a treatment or some arbitrary level of compliance with a medication that has been prescribed. The ICH E9 (R1) estimands framework provides an improved alternative for the description of treatment effects in clinical trials. Identification of important intercurrent events and the strategy used to handle these events is key to determining the treatment effect. When designing a trial, estimands should be properly defined according to this framework. It is time the statistical literature abandoned describing treatment effects as the effect in adherers or the per-protocol effect.

Treatment Detection and Movement Disorder Society-Unified Parkinson's Disease Rating Scale, Part III Estimation Using Finger Tapping Tasks.

The validation of objective and easy-to-implement biomarkers that can monitor the effects of fast-acting drugs among Parkinson's disease (PD) patients would benefit antiparkinsonian drug development. We developed composite biomarkers to detect levodopa/carbidopa effects and to estimate PD symptom severity. For this development, we trained machine learning algorithms to select the optimal combination of finger tapping task features to predict treatment effects and disease severity. Data were collected during a placebo-controlled, crossover study with 20 PD patients. The alternate index and middle finger tapping (IMFT), alternative index finger tapping (IFT), and thumb-index finger tapping (TIFT) tasks and the Movement Disorder Society-Unified Parkinson's Disease Rating Scale (MDS-UPDRS) III were performed during treatment. We trained classification algorithms to select features consisting of the MDS-UPDRS III item scores; the individual IMFT, IFT, and TIFT; and all three tapping tasks collectively to classify treatment effects. Furthermore, we trained regression algorithms to estimate the MDS-UPDRS III total score using the tapping task features individually and collectively. The IFT composite biomarker had the best classification performance (83.50% accuracy, 93.95% precision) and outperformed the MDS-UPDRS III composite biomarker (75.75% accuracy, 73.93% precision). It also achieved the best performance when the MDS-UPDRS III total score was estimated (mean absolute error: 7.87, Pearson's correlation: 0.69). We demonstrated that the IFT composite biomarker outperformed the combined tapping tasks and the MDS-UPDRS III composite biomarkers in detecting treatment effects. This provides evidence for adopting the IFT composite biomarker for detecting antiparkinsonian treatment effect in clinical trials. © 2023 The Authors. Movement Disorders published by Wiley Periodicals LLC on behalf of International Parkinson and Movement Disorder Society.

Application Value of Serum Neurofilament Light Protein for Disease Staging in Huntington's Disease.

Neurofilament light protein (NfL) has been proven to be a sensitive biomarker for Huntington's disease (HD). However, these studies did not include HD patients at advanced stages or with larger CAG repeats (>50), leading to a knowledge gap of the characteristics of NfL. Serum NfL (sNfL) levels were quantified using an ultrasensitive immunoassay. Participants were assessed by clinical scales and 7.0 T magnetic resonance imaging. Longitudinal samples and clinical data were obtained. Baseline samples were available from 110 controls, 90 premanifest HD (pre-HD) and 137 HD individuals. We found levels of sNfL significantly increased in HD compared to pre-HD and controls (both P < 0.0001). The increase rates of sNfL were differed by CAG repeat lengths. However, there was no difference in sNfL levels in manifest HD from early to late stages. In addition, sNfL levels were associated with cognitive measures in pre-HD and manifest HD group, respectively. The increased levels of sNfL were also closely related to microstructural changes in white matter. In the longitudinal analysis, baseline sNfL did not correlate with subsequent clinical function decline. Random forest analysis revealed that sNfL had good power for predicting disease onset. Although sNfL levels are independent of disease stages in manifest HD, it is still an optimal indicator for predicting disease onset and has potential use as a surrogate biomarker of treatment effect in clinical trials. © 2023 International Parkinson and Movement Disorder Society.

Artificial intelligence approach for the analysis of placebo-controlled clinical trials in major depressive disorders accounting for individual propensity to respond to placebo

Treatment effect in clinical trials for major depressive disorders (RCT) can be viewed as the resultant of treatment specific and non-specific effects. Baseline individual propensity to respond non-specifically to any treatment or intervention can be considered as a major non-specific confounding effect. The greater is the baseline propensity, the lower will be the chance to detect any treatment-specific effect. The statistical methodologies currently applied for analyzing RCTs doesn’t account for potential unbalance in the allocation of subjects to treatment arms due to heterogenous distributions of propensity. Hence, the groups to be compared may be imbalanced, and thus incomparable. Propensity weighting methodology was used to reduce baseline imbalances between arms. A randomized, double-blind, placebo controlled, three arms, parallel group, 8-week, fixed-dose study to evaluate efficacy of paroxetine CR 12.5 and 25 mg/day is presented as a cases study. An artificial intelligence model was developed to predict placebo response at week 8 in subjects assigned to placebo arm using changes from screening to baseline of individual Hamilton Depression Rating Scale items. This model was used to predict the probability to respond to placebo in each subject. The inverse of the probability was used as weight in the mixed-effects model applied to assess treatment effect. The analysis with and without propensity weight indicated that the weighted analysis provided an estimate of treatment effect and effect-size about twice larger than the non-weighted analysis. Propensity weighting provides an unbiased strategy to account for heterogeneous and uncontrolled placebo effect making patients’ data comparable across treatment arms.

Estimand in benefit-risk assessment

ABSTRACT ICH E9(R1) introduces the estimand framework to strengthen dialogues between sponsors and regulators during drug development. A well-structured benefit-risk assessment (BRA) framework also intends to facilitate communication among stakeholders. However, the estimand in ICH E9(R1) is written mainly from the perspective of a single measure of treatment effect in clinical trials. There is lack of systematic discussion on estimand in the context of BRA. This paper initiates the BRA discussion under the estimand framework. By identifying two types of BRA approaches, we summarize and discuss completed clinical trials, using the estimand language for BRA. Benefits and challenges of using estimand for BRA are also discussed.

The empirical estimate of the survival and variance using a weighted composite endpoint

BackgroundComposite endpoints for estimating treatment efficacy are routinely used in several therapeutic areas and have become complex in the number and types of component outcomes included. It is assumed that its components are of similar asperity and chronology between both treatment arms as well as uniform in magnitude of the treatment effect. However, these assumptions are rarely satisfied. Understanding this heterogeneity is important in developing a meaningful assessment of the treatment effect.MethodsWe developed the Weighted Composite Endpoint (WCE) method which uses weights derived from stakeholder values for each event type in the composite endpoint. The derivation for the product limit estimator and the variance of the estimate are presented. The method was then tested using data simulated from parameters based on a large cardiovascular trial. Variances from the estimated and traditional approach are compared through increasing sample size.ResultsThe WCE method used all of the events through follow-up and generated a multiple recurrent event survival. The treatment effect was measured as the difference in mean survivals between two treatment arms and corresponding 95% confidence interval, providing a less conservative estimate of survival and variance, giving a higher survival with a narrower confidence interval compared to the traditional time-to-first-event analysis.ConclusionsThe WCE method embraces the clinical texture of events types by incorporating stakeholder values as well as all events during follow-up. While the effective number of events is lower in the WCE analysis, the reduction in variance enhances the ability to detect a treatment effect in clinical trials.

Can We Use Blood Biomarkers as Entry Criteria and for Monitoring Drug Treatment Effects in Clinical Trials? A Report from the EU/US CTAD Task Force.

In randomized clinical trials (RCTs) for Alzheimer's Disease (AD), cerebrospinal fluid (CSF) and positron emission tomography (PET) biomarkers are currently used for the detection and monitoring of AD pathological features. The use of less resource-intensive plasma biomarkers could decrease the burden to study volunteers and limit costs and time for study enrollment. Blood-based markers (BBMs) could thus play an important role in improving the design and the conduct of RCTs on AD. It remains to be determined if the data available on BBMs are strong enough to replace CSF and PET biomarkers as entry criteria and monitoring tools in RCTs.

Plasma biomarkers predict amyloid pathology in cognitively normal monozygotic twins after 10 years.

Blood-based biomarkers could prove useful to predict Alzheimer's disease core pathologies in advance of clinical symptoms. Implementation of such biomarkers requires a solid understanding of their long-term dynamics and the contribution of confounding to their association with Alzheimer's disease pathology. Here we assess the value of plasma amyloid-β1-42/1-40, phosphorylated-tau181 and glial fibrillary acidic protein to detect early Alzheimer's disease pathology, accounting for confounding by genetic and early environmental factors. Participants were 200 monozygotic twins, aged ≥60 years with normal cognition from the european medical information framework for Alzheimer's disease study. All twins had amyloid-β status and plasma samples available at study enrolment. For 80 twins, additional plasma samples were available that had been collected approximately 10 years prior to amyloid-β status assessment. Single-molecule array assays were applied to measure amyloid-β1-42/1-40, phosphorylated-tau181 and glial fibrillary acidic protein. Predictive value of and longitudinal change in these biomarkers were assessed using receiver operating characteristic curve analysis and linear mixed models. Amyloid pathology could be predicted using blood-based biomarkers obtained at the time of amyloid status assessment (amyloid-β1-42/1-40: area under the curve = 0.65, P = 0.01; phosphorylated-tau181: area under the curve = 0.84, P < 0.001; glial fibrillary acidic protein: area under the curve = 0.74, P < 0.001), as well as using those obtained 10 years prior to amyloid status assessment (amyloid-β1-42/1-40: area under the curve = 0.69, P = 0.03; phosphorylated-tau181: area under the curve = 0.92, P < 0.001; glial fibrillary acidic protein: area under the curve = 0.84, P < 0.001). Longitudinally, amyloid-β1-42/1-40 levels decreased [β (SE) = -0.12 (0.01), P < 0.001] and phosphorylated-tau181 levels increased [β (SE) = 0.02 (0.01), P = 0.004]. Amyloid-β-positive individuals showed a steeper increase in phosphorylated-tau181 compared with amyloid-β-negative individuals [β (SE) = 0.06 (0.02), P = 0.004]. Also amyloid-β-positive individuals tended to show a steeper increase in glial fibrillary acidic protein [β (SE) = 0.04 (0.02), P = 0.07]. Within monozygotic twin pairs, those with higher plasma phosphorylated-tau181 and lower amyloid-β1-42/1-40 levels were more likely to be amyloid-β positive [β (SE) = 0.95 (0.26), P < 0.001; β (SE) = -0.28 (0.14), P < 0.05] indicating minimal contribution of confounding by genetic and early environmental factors. Our data support the use of amyloid-β1-42/1-40, phosphorylated-tau181 and glial fibrillary acidic protein as screening tools for Alzheimer's disease pathology in the normal aging population, which is of importance for enrolment of high-risk subjects in secondary, or even primary, prevention trials. Furthermore, these markers show potential as low-invasive monitoring tool of disease progression and possibly treatment effects in clinical trials.

Investigating neurodegenerative heterogeneity in cognitively unimpaired A4 pre‐randomization cohort

AbstractBackgroundThe heterogeneous nature of Alzheimer’s disease and uncertainty of determining disease severity and subtype can confound treatment effects in clinical trials. Computational methods for subtyping can assist in trial enrichment. This study investigates the latent heterogeneity and severity (including absence) of neurodegeneration in the cognitively unimpaired A4 cohort enriched for elevated amyloid.MethodBaseline 3T T1‐weighted MRI data from ADNI (N=687) and A4 pre‐randomization (N=1240, all Aβ+ [Sperling et al. 2020, DOI:10.1001/jamaneurol.2020.0387]) studies were included. FreeSurfer was used to estimate regional brain volumes, of which 13 features were input to SuStaIn (Young et al., 2018, DOI:10.1038/s41467‐018‐05892‐0) to train two cohort‐specific models of neurodegeneration subtypes. Volumes were first adjusted for age, sex, education, and intracranial volume, and z‐scored relative to controls (Aβ– in A4; cognitively normal in ADNI). For each feature, z‐score events of z=1 (subtle abnormality) and z=2 (atrophy) were used. The number of subtypes was determined using cross‐validation.ResultFigure 1 shows our SuStaIn model for ADNI, replicating the three‐subtype finding of Young et al. (2018). Figure 2 shows our SuStaIn model for A4, which estimated a similar three‐subtype model: typical, cortical, subcortical (named for the earliest atrophy events). Notable differences between the models are in the subtle abnormality (z=1) events in A4 that might, or might not, develop into more prominent atrophy (z=2), such as with the caudate which stalls at z=1 (Figure 2, top‐left). Our results indicate the presence of AD‐like heterogeneity in the A4 cohort. Figure 3 shows boxplots of amyloid PET SUVr — there are no statistically significant differences between subtypes (Mann‐Whitney U test), suggesting amyloid PET screening is insufficient to detect neurodegeneration heterogeneity. There were no statistically significant subtype differences in cognitive measures nor demographics (Table 1).ConclusionOur results show that neurodegeneration heterogeneity exists in the A4 cohort, despite amyloid PET screening. This heterogeneity could have ramifications on the potential success of the A4 trial, particularly if it correlates with different rates of cognitive decline and confounds anti‐amyloid therapeutic effects on trial endpoints. Our results support the use of more holistic, multimodal, data‐driven screening for clinical trials in Alzheimer’s disease to reduce cohort heterogeneity.