Original Title: Efectividad y coste-efectividad de la determinación de las proteínas abeta-42, tau total y tau fosforilada en líquido cefalorraquídeo como biomarcadores de la enfermedad de Alzheimer

Authors' objectives: GENERAL AIM: To determine the effectiveness and cost-effectiveness of the Aβ1-42, T-tau and P-tau proteins in CSF as biomarkers of AD. SPECIFIC OBJECTIVES: O.1. To determine the diagnostic performance of biomarkers in CSF (Aβ1-42, T-tau, and P-tau proteins) for the diagnosis of AD, in studies with control groups. O.2. To determine the efficacy or diagnostic performance of biomarkers in CSF (Aβ1-42, T-tau, and P-tau proteins) with the postmortem definitive diagnostic of AD. O.3. To determine the diagnostic performance of biomarkers in CSF (Aβ1-42, T-tau, and P-tau proteins) to predict the risk of progress from mild cognitive impairment (MCI) to AD. O.4. To determine the diagnostic performance of biomarkers in CSF (Aβ1-42, T-tau, and P-tau proteins) for the differential diagnosis of AD with other types of dementia: Lewy body dementia (DCLewy), frontotemporal lobe degeneration (FTLD), vascular dementia (VD) and Creutzfeldt-Jakob disease (CJD). O.5. To determine the efficiency (cost-effectiveness) of the determination of Aβ1-42, T-tau, and P-tau proteins in CSF like biomarkers for early detection of AD and confirmation of the diagnosis.

Authors' results and conclusions: RESULTS: 1. Diagnostic performance. 1.1. Systematic review with meta-analysis. From the search strategy focused on systematic reviews with metaanalysis, we obtained a total of 1770 references, of which 63 references were selected by title and abstract, and finally 7 were included: Bloudek et al. (2011), Diniz et al. (2008), Mitchell (2009), Monge-Argilés et al. (2010), Schmand et al. (2010), Sunderland et al. (2003), and van Harten et al. (2011). These meta-analysis include 130 references non repeated, of which 72 compared AD vs. healthy controls (Aβ1-42: 30, T-tau: 50; Ptau: 26), 23 compared people with mild cognitive impairment-converter vs. people with mild cognitive impairment-stable (Aβ1-42: 13, T-tau: 20, P-tau: 15), and 78 references analyzed the discriminative power between AD and other non AD dementias (Aβ1-42: 11, T-tau: 60; p -tau: 41). 1.2. Primary diagnostic studies. From the strategy employed for the primary studies 1284 references were obtained, of which 352 were selected based on title and abstract. If several studies evaluated the same sample, the study with the largest sample size was included. At the end of the process, a total of 54 articles were included in the review (22 cross-sectional and 32 longitudinal studies), of which 32 were considered for meta-analyzes. The studies have very diverse follow-up, ranging from 6 months to 6.8 years. 2. ECONOMIC EVALUATION: Results in Scenario 1 demonstrate the potential cost-effectiveness of biomarkers for the diagnosis of MCI related to AD. For this group of patients an ICER of € -451,65 [€ -682,06; € 1000,34] for correct diagnosis is obtained. With these results, it can be stated that the use of biomarkers of AD as a diagnostic technique is a cost-effective alternative for willingness to pay (WTP) ranging from 20000 € to 2000 € for correct diagnosis. The acceptability curve in this scenario provides very low error probabilities for the WTP mentioned, when the alternative of CSF biomarkers is accepted as cost-effective versus diagnostic criteria. The parameters that have the most influence on the model developed are sensitivity of the diagnostic criteria and specificity of biomarkers in the diagnosis of AD. In Scenario 2, the uncertainty about the model parameters is high. In this group of patients, there is no evidence that the alternative of biomarkers is more cost-effective than standard diagnostic criteria in terms of diagnostic efficiency. In this case the ICER is € 411,57 and its confidence interval covers the four quadrants of the cost-effectiveness plane, which prevent us from determining which technique is more efficient. The parameters that influence the most on this second group of patients are specificity of biomarkers and the cost of the lumbar puncture. CONCLUSIONS: 1. DIAGNOSTIC PERFORMANCE: A clear pattern of superiority of a CSF biomarker or a combination of them has not been identified, with the sensitivity and specificity levels achieved around 80%, when comparing the results between patients with AD and controls without dementia. There is quantitatively and qualitatively sufficient evidence showing that lower levels of Aβ1-42 and increased T-tau and phosphorylated P-tau levels are a specific feature of AD. The validity of the combined use of CSF biomarkers finds its best diagnostic performance to predict progression from MCI to AD, with appropriate sensitivity level and lower specificity level. This evidence, despite its consistency and robustness, is limited to only three studies. 32 INFORMES, ESTUDIOS E INVESTIGACIÓN For differential diagnosis of AD versus other dementias together, the ratio P-tau/Aβ1-42 has shown the most diagnostic efficacy. Studies that have compared AD versus other causes of dementia have generally observed sensitivity levels for T-tau and P-tau not higher than 80%. 2. ECONOMIC EVALUATION: The determination of Aβ-42, T-tau, and P-tau proteins is a cost-effective alternative in the diagnosis of MCI related to AD, from the perspective of the NHS if WTP is asumed. However, since there is no defined WTP in our country to this unit of effectiveness, these results should be treated with caution. It cannot be concluded from the results of this EE that the use of biomarkers in CSF as a diagnostic alternative is more costeffective than standard diagnostic criteria in patients with clinical diagnosis of dementia. More research is needed to assess the impact of the introduction of this technology. In the short-medium term, the impact in the clinical management of patients and the amount of resources needed should be assessed. In the long term, should be considered the early treatment outcomes in patients with MCI related to AD. This information is necessary for economic evaluations that inform about the influence of such diagnoses in health outcomes (in terms of QALYs or LY) patients. Despite favorable evidence on the diagnostic performance of CSF biomarkers for AD, it is necessary to standardize, standardize and formalize the different biomarkers in CSF for the diagnosis of AD, given the wide variation between studies.

Authors' recommendations: If WTP thresholds considered in this economic evaluation would be assumed, we recommend the determination of proteins Aβ1-42, T-tau, and P-tau in CSF as a cost-effective alternative for the diagnosis of MCI related to AD. From the results of the economic evaluation, there is insufficient evidence to recommend the use of biomarkers in patients with dementia.

Authors' methods: EVALUATION OF DIAGNOSTIC PERFORMANCE: A systematic review (SR) was conducted, composed by two complementary search strategies, to evaluate the diagnostic performance of Aβ1-42, T-tau, and P-tau proteins in CSF as biomarkers of AD (sensitivity, specificity, positive predictive, negative predictive value). The first electronic search was conducted without time limit (until September 2012) to identify SR with meta-analysis (MA). The second search was focused on primary studies published from 2002 to September 2012. We also reviewed economic evaluations, clinical practice guidelines, and relevant consensus of experts on the subject. The process of study selection was performed by 8 reviewers. The peer review was done independently, and in case of doubt and/or disagreements a third reviewer was consulted. The literature review was supplemented by manual references extracted from the included studies, if they met the selection criteria. The critical evaluation and the data extraction of the references included also was held by 8 reviewers, who analyzed the quality and risk of bias of the studies by independent peers. The MA was assessed based on the Oxman scale, and the primary studies based on QUADAS2 scale. For each objective of this work, additional MA of diagnostic performance were conducted with those studies that provided data on sensitivity and specificity, or where it was possible to calculate, for each biomarker studied or the ratios between Aβ1-42 and one of the other two markers. ECONOMIC EVALUATION: A Complete Economic Evaluation (EE) was performed by means of a cost-effectiveness analysis (CEA) which compared two diagnostic alternatives for AD the determination of abeta-42 proteins, total tau and phosphorylated tau in CSF as biomarkers of AD and the standard diagnostic criteria (NINCDS-ADRDA); because no data are available on the sensitivity and specificity of the new NIA criteria. The effectiveness measure used was "diagnosed and treated properly" and analysis was performed from the perspective of the NHS. Thus only direct healthcare costs were included. The model developed for the CEA consists of a decision tree with two main branches representing each of the compared diagnostic strategies. We analyzed the results from two scenarios with patients of different categories. In the first scenario (Scenario 1), a hypothetical cohort of patients 60 years or older is assumed, with MCI at the time of diagnosis. In the second scenario (Scenario 2), a hypothetical cohort of patients 60 years or older is assumed, with symptoms of dementia at the time of undergo a diagnostic alternative. The estimation of the model parameters was performed from the literature. Data from observational studies and economic evaluations were combined with expert opinions on each of the two diagnostic alternatives under study. A multivariate and probabilistic analysis was applied by means of Monte Carlo simulation of 2nd order in which the probability distributions that best fit the behavior of the parameters were used. For each of the scenarios considered 10.000 Monte Carlo simulations of 2nd order were performed and pairs cost-effective solutions in the incremental costeffectiveness plane were plotted. Subsequently, acceptability curves for each of the scenarios were calculated and ANCOVA models were applied to the results of Monte Carlo simulations in order to identify those parameters that caused greater variability in the model results. Analyses were performed with TreeAge Pro 2009 software Healthcare (TreeAge Software, Inc., USA) and the statistical package STATA 11.

Project Status: Completed

Year Published: 2012

URL for published report: https://sescs.es/biomarcadores-de-la-enfermedad-de-alzheimer/

English language abstract: An English language summary is available

Publication Type: Full HTA

Country: Spain

Organisation Name: Canary Health Service

Contact Address: Dirección del Servicio. Servicio Canario de la Salud, Camino Candelaria 44, 1ª planta, 38109 El Rosario, Santa Cruz de Tenerife

Contact Name: sescs@sescs.es

Contact Email: sescs@sescs.es