Biological Variation of Cardiac Troponins in Health and Disease: A Systematic Review and Meta-analysis

• Published in: Clinical chemistry (Baltimore, Md.) Vol. 67; no. 1; pp. 256 - 264
• Main Authors: Diaz-Garzon, Jorge, Fernandez-Calle, Pilar, Sandberg, Sverre, Özcürümez, Mustafa, Bartlett, William A, Coskun, Abdurrahman, Carobene, Anna, Perich, Carmen, Simon, Margarita, Marques, Fernando, Boned, Beatriz, Gonzalez-Lao, Elisabet, Braga, Federica, Aarsand, Aasne K
• Format: Journal Article
• Language: English
• Published: England Oxford University Press 01.01.2021; American Association for Clinical Chemistry, Inc
• Subjects: Biological variation; Biological Variation, Individual; Biomarkers - analysis; Calcium-binding protein; Cardiovascular Diseases - diagnosis; Confidence intervals; Estimates; Heart diseases; Humans; Kidney Diseases - diagnosis; Medical research; Medicine, Experimental; Meta-analysis; Prognosis; Reference Values; Sampling; Troponin; Troponin I; Troponin I - analysis; Troponin I - standards; Troponin T; Troponin T - analysis; Troponin T - standards; Variation; Spain; meta-analysis; Biological variation; systematic review; BIVAC; cardiac troponin
• ISSN: 0009-9147; 1530-8561; 1530-8561
• DOI: 10.1093/clinchem/hvaa261

Abstract Background Many studies have assessed the biological variation (BV) of cardiac-specific troponins (cTn), reporting widely varying within-subject BV (CVI) estimates. The aim of this study was to provide meta-analysis-derived BV estimates for troponin I (cTnI) and troponin T (cTnT) for different sampling intervals and states of health. Methods Relevant studies were identified by a systematic literature search. Studies were classified according to their methodological quality by the Biological Variation Data Critical Appraisal Checklist (BIVAC). Meta-analyses of BIVAC-compliant studies were performed after stratification by cTn isoform, exclusion of results below the limit of detection, states of health, and sampling interval to deliver reference change values (RCV), index of individuality (II) and analytical performance specifications (APS) for these settings. Results Sixteen and 15 studies were identified for cTnI and cTnT, respectively, out of which 6 received a BIVAC grade A. Five studies had applied contemporary cTnI assays, but none contemporary cTnT. High-sensitivity (hs-) cTnI and cTnT delivered similar estimates in all settings. Long-term CVI estimates (15.1; 11.3%) derived from healthy individuals were higher than short-term (4.3%; 5.3%) for hs-cTnI and hs-cTnT, respectively, although confidence intervals overlapped. Estimates derived from diseased subjects were similar to estimates in healthy individuals for all settings. Conclusions This study provides robust estimates for hs-cTnI and hs-cTnT applicable for different clinical settings and states of health, allowing for the use of RCV both to aid in the diagnosis of myocardial injury and for prognosis. BV-based APS appear too strict for some currently available technologies.