PRE-ANALYTICAL VARIABLES IN THE CONTEXT OF PUBLISHING/ESTABLISHING A STANDARDIZED PRE-ANALYTICAL PROTOCOL FOR ALZHEIMER’S DISEASE (AD) RESEARCH USE DANNI LI, SILVIA FOSSATI, DOUGLAS GALASKO AND MICHELLE MIELKE, ON BEHALF OF THE BBB PIA REFERENCE RANGES WORKING GROUP GBSC AND SABB JOINT MEETING 10/30/18
BACKGROUND ON THE CHECKLIST COMPILED BY THE BIOFLUID BIOMARKERS PIA REFERENCE RANGES WORKING GROUP • Reference ranges depend on many things: pre-analytical factors, assay methods, intended use, patient populations etc. • In order to establish ref ranges, we need to have a comprehensive understanding of all these factors (including pre-analytical factors) and how they vary from studies to studies.
REFERENCE RANGES WORKING GROUP • Chairperson/Facilitator: Dr. Michelle Mielke • 22 members worldwide • Initial goals/steps 1. Identify and decide on a few promising biofluids based markers that may have future use at the general physician level 2. Compare/contrast study sample collection; assays; populations (e.g., exclusion criteria) 3. While ultimate goal may be ‘normal ranges’, differences in study methods and clinical assays will be a factor 4. Decide on how best to move forward (e.g., standardize and re-assay measures?) 5. Identify studies interested in donating samples to this effort and participating
BBB PIA REFERENCE RANGES WORKING GROUP DEVELOPED AND TESTED A COMPREHENSIVE CHECKLIST Checklist components: Tested using 4 recently Sample collection (pre- • published manuscripts on analytical variables) blood amyloid peptides. Assay information • This check list was (analytical) shared with the joint meeting members. Assay information • (clinical) and statistical analysis Study population •
Teunissen et al 2018 Ovod et al 2018 AlzDem Lovhem et al 2017 AlzDem Nakamura et al 2018 Sample collection (pre- JAD (Michelle) (Danni Li) (Silvia) Nature Letter (Doug) analytical variables) NOTE: preanalytical factors may matter because data from Pre-analytical Japanese subjects and AIBL subjects had different dynamic variables ranges. N/A Approximately hourly in 24 Early morning (80%) morning Time of collection hours. These 13 variables were N/A Fasting over night (although Yes. 4h fasting (20%) or unknown Fasting status blood samples were overnight fasting (80%). The selected based on literature, collected hourly for a 20 authors previously evaluated the time points after IV bolus or effects of fasting with no consensus, and the group’s oral intake of 800 mg significant variations in plasma clinical chemistry expertise. labelled Leucine) ab40and ab42 (Hansson 2010) N/A N/A Yes If non fasting, time from last meal N/A N/A N/A Needle size and location of draw N/A Yes (samples were N/A Handling of tubes Guidelines for the standardization of centrifuged immediately on preanalytic variables for blood-based collection). EDTA Polypropylene tubes EDTA (however this is not clearly EDTA biomarker studies in Alzheimer's Tube types and additives, (Axygen) specified in the referenced paper e.g., EDTA, acid citrate, disease research. (Hallmans 2003) heparin O'Bryant SE, et al. N/A N/A N/A Tube collection order Alzheimers Dement. 2015 N/A I am not sure what this N/A Time of sample in May;11(5):549-60. doi: means. It is very likely this collection tube 10.1016/j.jalz.2014.08.099. Epub information is not provided. 1500-2500 g for 15 N/A N/A Centrifugation parameters 2014 Oct 1. Review. minutes N/A N/A N/A N/A Time from collection to freeze -80C negative 80 celsius negative 70 celsius -80C Temperature of freeze 1 No explicit information is N/A Freeze-thaw cyles provided (maybe never thawed samples?) 0.5 ml N/A 1.5mL unknown Aliquot size
Assays (including analytical performance) and biochemical testing process Methodology (e.g., antibody based immunoassay, mass spectrometry based) and platform if relevant (e.g., MSD) If ELISA or similar: chemiluminescent/ fluorescent/ PCR readout Validated for multiplexing or not Commericially available or home brewed Quantitative (e.g., calibration curve established to derive quantitative results) Quality control samples used Assay Calibration standards used e.g., recombinant peptides (including source/vendor) Compatible additives (EDTA, heparin etc) Methodology and Compatible biofluid (plasma or serum) Between day precisions biochemical Assay sensitivity Assay specificity process variables Volume of sample needed Is sample diluted? What diluent? LLOD ULOD LOQ Linear range Heterophile Abs cause interference? Intralab CV for assay (intra-assay) Intralab CV for assay (inter-assay) Inter lab CV for assay Samples tested randomized? Samples tested blinded? Samples tested in multiple batches? If in batches, did each batch contain equal N for each study group? Reagent lot changed during the biochemical testing process? Dilution experiments performed during assay validation Spike-in experiments performed Detergent, protease inhibitor or other reagent added Immunoprecipitation protocol
Assays (including clinical performance) and statistical analysis Intended use or context of use Dependent variable Assay clinical Independent variable: AD related outcomes (e.g.,clinical diagnosis, performance and neuroimaging data, or other fluid biomakers) Statistical analysis Clinical sensitivity Variables Clinical specificity Positive predictive value Negative predictive value Effect size Brief discrption of the statistical method used Has the results been validated in other independent samples? If so, were the assay conducted by a different lab?
Study population Sample size Age Sex Study Population Ethinicity Variables Race APOE e4 Smoking status (current/former/never) Gestation Diet Medications (e.g., lipid lowering, antihypertensive) No-AD comorbidiites: cardiovascular diseases, diabetes, cerebrovascular disease (e.g., stroke), hypertension, psychiatric disorders, TBI Alchohol use: current/former/never Physical activity level Exclusion criteria used BMI Renal function Diabetes
Insights gained from this exercise that may be helpful for a pre-analytical protocol discussion: #1: ORIGINAL RESEARCH STUDIES ARE NOT DESIGNED TO STUDY PRE-ANALYTICAL/ANALYTICAL VARIABLES • Comprehensive pre-analytical variables information are lacking from original research studies (e.g., 4 references included in our checklist exercise). • Purpose of these original research studies are not to evaluate pre-analytical or analytical variables. • Studies of pre-analytical/analytical variables require a different study design.
EXAMPLES OF STUDIES DESIGNED TO EVALUATE PRE-ANALYTICAL VARIABLES
#1B. STUDY RESULTS ARE DIFFERENT DUE TO MANY FACTORS, NOT JUST IN HOW SAMPLES ARE COLLECTED (EG. PRE- ANALYTICAL VARIABLES), BUT ASSAY METHODS, PATIENT POPULATIONS ETC. • If interested in assay method and performance difference, do a round robin study as Kaj/Henrik suggested using the same set of samples collected under the most pristine condition possible (eliminate pre-analytical variables) to be measured by different methods • If interested in pre-analytical variables, do a preanalytical variable study (responsibility of research group, a few groups or assay manufacturer) since these preanalytical variables may be method specific. • If interested in comparing clinical performance of specific context of use, do a round robin study with the same set of samples collected under the most pristine condition possible with clinical information.
#2: PRE-ANALYTICAL VARIABLES CAN BE METHOD (OR TARGET) SPECIFIC. • This does not mean that a generic pre-analytical protocol for all AD biomarkers is not possible. • However, studies of pre-analytical variables should include all intended methods/biomarkers. • A clarifying question could be: will the standardized protocol be applicable to all promising AD blood biomarkers/methods? • Amyloid peptides (Simoa, IMR, IP-MS, and etc) • NFL (Simoa, IMR and etc) • Total Tau (Simoa, IMR and etc) • P-tau (Simoa, IMR and etc)
TOP BIOMARKER CANDIDATES (FOR RR WORKING GROUP) We received a total of 29 votes (up to 3 candidates per member) • NfL (9) • Screening/diagnosis, prognosis, progression, and treatment efficacy • p-tau 181 (5), total tau (4), p-tau 231 (1), p-tau 396 (1) • Screening/diagnosis, prognosis, progression, and treatment efficacy • Amyloid beta 42, 40, 38 or ratio (4) • Screening/diagnosis and treatment efficacy
QUESTIONS • Should the standardized protocols be analyte and/or method specific? • If so, what blood analytes and methods should these standardized protocols include? • Can we establish best pre-analytical protocols from the literature or do we need to test them in new studies?
Recommend
More recommend
