https://orcid.org/0009-0009-8296-4076
Figures
Abstract
Background
Sepsis is a common admission diagnosis in the intensive care unit (ICU). The Sepsis-3 consensus associates sepsis diagnosis with acute organ dysfunction. In these patients troponin elevation is a well-established phenomenon, but its clinical significance is not settled, as no systematic review has addressed the prognostic significance of the increasingly prevalent high-sensitivity troponin assays in acute organ dysfunction setting.
This study aims to clarify the association between early serum troponin levels in high-sensitivity assays with short-term mortality risk in septic patients with acute organ dysfunction.
Methods
We will systematically search PubMed, Scopus and Embase for original articles; additionally, a manual search will be carried out through relevant literature. Generally, studies will be deemed eligible for inclusion if they evaluate the association between high-sensitivity troponin in the first 24 hours of admission and ICU, 30-days, or In-hospital mortality; in patients with septic shock or sepsis related to acute organ dysfunction. Two reviewers will independently select studies and extract the data. A meta-analysis for mortality outcome will be performed for comparative data regarding two effect measures: Odd ratios and Standardized Mean differences.
Discussion
This study will provide further evidence about the role of high-sensitivity troponin assays in predicting mortality in septic patients; potentially helping to guide further research and yielding valuable information for patient assessment.
Conclusion about the certainty of evidence will be presented in a ´Summary of findings´ table.
Citation: Ferrière-Steinert S, Valenzuela Jiménez J, Heskia Araya S, Kouyoumdjian T, Ramos-Rojas J, Gajardo AIJ (2024) Early high-sensitivity troponin elevation in predicting short-term mortality in sepsis: A protocol for a systematic review with meta-analysis. PLoS ONE 19(10): e0301948. https://doi.org/10.1371/journal.pone.0301948
Editor: Sonu Bhaskar, Global Health Neurology Lab / NSW Brain Clot Bank, NSW Health Pathology / Liverpool Hospital and South West Sydney Local Health District / Neurovascular Imaging Lab, Clinical Sciences Stream, Ingham Institute, AUSTRALIA
Received: March 25, 2024; Accepted: August 16, 2024; Published: October 25, 2024
Copyright: © 2024 Ferrière-Steinert et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: No datasets were generated or analysed during the current study. All relevant data from this study will be made available upon study completion.
Funding: This study was funded by the Government of Chile, through the grant: Initiation in research FONDECYT grant 11241548, National Agency of Research and Development (https://anid.cl/about-us/); Awarded to Abraham I. J. Gajardo. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
1. Background
Sepsis is the most common mortality cause in Intensive Care Unit (ICU) patients in the United States. A fourth of patients diagnosed with sepsis will die within their hospitalization, rising to a half within the septic shock group [1]. Furthermore, sepsis-related deaths have been on the rise in the US and worldwide [1–3], and are expected to continue to do so.
Sepsis definitions have varied throughout the years. Sepsis-3 consensus, agreed in 2016, defines sepsis as a life-threatening organ dysfunction (assessed by SOFA score) in behalf of an infection; disposing of the former category of severe sepsis while preserving septic shock [4]. Thus, compared with previous sepsis definitions, patients diagnosed with sepsis-3 criteria are exposed to an increased mortality and adverse events risk [5].
Sepsis is associated with complications and dysfunction of several systems [6]. At the cardiac level, sepsis is well-documented to be associated with acute myocardial injury independent of coronary perfusion abnormality [7]. This phenomenon is primarily attributed to inflammatory and cardio-depressant factors that act indirectly and directly on the cardiomyocyte [7,8]. Furthermore, when the myocardial injury ensues in diastolic and/or systolic dysfunction, it is considered Septic Cardiomyopathy [7]. Although there is no consensus on the criteria for this entity, commonly accepted cardinal features are: The absence of an underlying coronary etiology, the acute presentation, and the reversibility of the dysfunction [7–9].
Cardiac troponins are widely used in the diagnosis of acute coronary syndrome and other cardiac pathologies presenting with myocardial injury [10], including sepsis. High-sensitivity cardiac troponin assays (hs-cTn) are usually defined by their capacity to detect troponin in at least 50% of healthy individuals [10,11]. In recent years hs-cTn have increasingly replaced conventional assays since they have both a lower limit of detection, and a lower and categorically defined normality cut-off. Thus, a hs-cTn level above the 99th percentile for the healthy population allows for a precise differentiation between normal and elevated groups [11].
Troponin elevation in septic patients is well-established but debated as a prognostic risk factor for clinical outcomes [7,12]. Three meta-analyses have been published on the issue [13–15], all showed a significant association of troponin elevation with increased mortality, but suffer from many problems of published data on troponin and mortality in sepsis: Lack of strict or similar criteria for sepsis diagnosis, heterogeneity of sampling time, usage of conventional rather than hs-cTn essays, and absence of controlling of confounders [7]. Moreover, the emergence and consolidation of Sepsis-3 definition, and the rise of hs-cTn assays could impact on the stratification of patients, modifying the prognostic association of the biomarkers with mortality risk [16]. Consequently, it is currently unknown if elevation in hs-cTn levels is associated with mortality risk in patients with sepsis by Sepsis-3 consensus.
This study aims to assess the association of early high-sensitivity serum troponin levels with short-term mortality in patients admitted to ICU because of sepsis with acute organ dysfunction.
2 Methods
2.1. Registration
The protocol was registered on the International Prospective Register of Systematic reviews (CRD42024468883). This protocol aligns with Preferred Reporting Items for Systematic review and Meta-Analysis Protocols (PRISMA-P) statement [17], and follows Riley´s guide for systematic review of prognostic factors [19] (see S1 Checklist). Deviations from protocol will be recognized in the final article.
2.2. Literature search
With the help of a knowledgeable librarian a search strategy was created: Pubmed, Scopus and Embase will be searched for primary articles published in any language and involving humans, with no date restrictions. The prepared search strings combine free text and MeSH/EMTREE terms for all words related to troponin and sepsis (see Table 1). No additional filters besides the ones described in the strings will be utilized.
Subsequently, we will manually scrutinize references of included studies and previous reviews on the topic to retrieve missing pertinent papers.
Searches will be re-run just before the final analyses to check for new studies that meet the inclusion criteria.
2.3. Study selection
In accordance to CHARMS guidance [18], the research question was formulated with the Population/Index prognostic factor/Comparator prognostic factors/Outcome/Timing/Setting (PICOTS) system, a modified version of the traditional PICO, where the P and O remains the same as the original, but the I stands for Index prognostic factors and C for other prognostic factors that can be considered as Comparators [19]. (See Table 2 for details on the defined scope of the systematic review). For the determination of inclusion and exclusion criteria Table 3 was constructed.
Based on this criteria, after deleting duplicates the corpus of studies will be scanned in Rayyan [20] by two independent reviewers, discrepancies will be solved by a third reviewer. Studies will be discarded based on 1 criterion only.
Initially, studies will undergo screening based on their title and abstract. Papers that meet the criteria during this preliminary assessment, or in cases where such analysis appears insufficient, will undergo further evaluation through full text assessment by the same reviewers for final selection.
Basically, interventional and observational primary studies will be deemed eligible if they reported some form of association between troponin levels and mortality in a general clinical context. Interventional studies will be handled as observational cohorts, extracting associations for both arms when previous evidence suggests that the intervention is not directly confounding the relation between troponin and mortality; if such a confounding bias mechanism is deemed plausible, only the placebo arm data will be utilized. We will only include studies that report on patients diagnosed either with sepsis by Sepsis-3 criteria [4], with Severe sepsis by Sepsis-1 or Sepsis-2 [21,22], or Septic shock by any of them [4,21,22]. Selection will be restricted to studies that specify the utilization of any hs-cTn assay in the first 24 hours after hospital admission, or whose reported limit of detection we consider high-sensitivity (i.e. < 9 ng/L)[10]. In the case that the latter information about the assays is omitted, we will exclude studies if they fail to meet a normality cutoff of 20 ng/L or lower, since they will be regarded as having a cutoff over the 99th percentile of the healthy population and therefore considered as non hs-cTn assay, based on European Society of Cardiology guidelines [10].
The study selection process will be reported using the PRISMA flowchart, providing a clear visual representation of the inclusion and exclusion criteria application [17].
2.4. Data extraction
Based on CHARMS-PF checklist [18] we designed a standardized data extraction form, containing the relevant characteristics of studies to be included (Table 4). Data will be extracted by two reviewers. Types of reported associations suitable to be extracted for this table are: 1) “dose-response”: data reported as odds, risk, or hazard ratio per unit increase in exposure; 2) “category/quantile based”: numbers (ideally 2x2 tables) or ratios comparing groups as defined by quantiles or categories of exposure; 3) “means”: data reported as means or mean differences in exposure, comparing those presenting and not presenting the event [23]. If a study presents more than one of these associations, we will extract all of them.
We will collect extracted effect measures together with their standard errors (SE). If SEs are not explicitly available, we will derive them from metrics such SDs, exact p-values, or confidence intervals [24]. We will collect adjusted and unadjusted effect measures separately. If many adjusted models are presented, we will extract the one which fits better the core adjustment factors selected (see Risk of Bias section). If effect measures for several time points are presented, we will prioritize ICU and 28-mortality (in that order) over in-hospital mortality. If effect measures for several hs-cTn samples are presented, we will prioritize the sample extracted closest to ED/ICU admission. Data only presented as figures will be extracted with WebPlotDigitizer software [25].
Authors will be contacted to request critical (i.e. data mandatory for PICOTS assessment) missing or not reported data regarding associations of interest. In case of no response, the article will be excluded from this study.
We will present a summarized version of this table in the final review.
2.5. Risk of bias assessment
Bearing in mind the prognostic nature of the review and taking into account previously published meta-analysis [13–15], we can expect that mainly observational studies will be included.
Two authors will independently assess the risk of bias (RoB) applying the QUIPS (quality in prognostic factor studies) tool [26], and then discuss together their assessments in order to decide the final score. In QUIPS RoB assessments are made within a set of 6 domains: (1] Study Participation, (2) Study Attrition, (3) Prognostic Factor Measurement, (4) Outcome Measurement, (5) Study Confounding and (6) Statistical Analysis and Reporting. We defined Study participation, Prognostic Factor Measurement and Study Confounding as the key domains for our assessment, and we specified criteria for each of these domains. For the confounder domain RoB evaluation, we will look for 3 core adjustment factors in each study: severity of sepsis, age, and comorbidities (any cardiac or renal).
Classification for each RoB item could be ‘low’, ‘moderate’, or ‘high’. Moreover, this assessment will include an overall judgment concerning the total RoB quality of each study, which will be visualized in the review.
2.6. Statistical analysis
We will investigate the mortality outcome with two effect measures: Mortality risk and Mean difference in troponin distribution.
1) Regarding mortality risk analysis: Whenever feasible a 2x2 contingency table will be gathered from available data, with one axis for survivor vs non-survivor groups, and other for elevated vs normal troponin level group, according to the cut-off defined by every study for each specific hs-cTn assay. If the sample is divided into more than 2 groups or categories (e.g. quantiles) it will be dichotomized around the closest match to a normality cutoff of 15 ng/L [27]. We will add 0.5 to each cell in any table that contains one or more zero values [24]. We will compute an Odds Ratio (OR) with its SE from this contingency table for each study. For studies not presenting sufficient data for a contingency table we will extract explicitly reported unadjusted ORs with their SEs. ORs from studies handling hs-cTn as a continuous variable (i.e. dose-response) will be scaled to an OR per 50 ng/L of increment. This magnitude was considered representative of the difference between the group usually referred to as “normal” compared to the “elevated” troponin level group–based on the reported troponin distributions of pertinent studies known to us and in accordance with our clinical expert [28–30].
Results from studies reporting Hazard Ratio (HR) or Relative Risk (RR) will be presented separately. Only if the event probability in the control group is <0.2, and the HR or RR are below 2.5, the effect measures will be interpreted as numerically equivalent to an OR and reported together with the other studies [31].
With these effect measures we will perform a structured quantitative analysis (meta-analysis) for the pooled unadjusted mortality effect.
ORs of adjusted (i.e. multivariate) regressions will be included in a separate model for the adjusted mortality effect regardless of the set of confounders controlled for. We will report variables incorporated in the regression for each study in a table.
2) For difference in troponin levels analysis: Studies presenting mean difference data in troponin level by survivor status will be meta-analyzed in a single model using Hedges´ g Standardized Mean Differences (SMD).
All models will be generated using R-4.3.1 package metafor [32], utilizing the random-effects model and the inverse variance weighting method [24].
Heterogeneity will be assessed and reported by I2 statistic, where >50% and a p-value < 0.10 are of concern and will be further explored [24].
Studies not presenting minimum data (e.g. missing SE) will be included in the narrative synthesis and compared to the results of the meta-analysis to the extent possible.
For the sake of replicability all formulas, raw data, code, calculations, model specifications and estimates shall be included either in the article itself or an online supplement.
2.7. Subgroup and sensitivity analysis
Models will undergo a comprehensive heterogeneity assessment. Since reviews of prognostic studies usually can result in high I2 regardless of similar point estimates [33], we will favor visual inspection over statistical criteria.
If in any of the model’s heterogeneity is deemed substantial, we will run the following Subgroup analysis:
- Subgroups by shock status. To assess whether effect measures and heterogeneity is significantly dependent on the degree of organic dysfunction, we will proceed with a subgroup analysis between studies restricting inclusion criteria to septic shock only, versus studies that encompass patients with sepsis, severe sepsis and septic shock.
- Subgroup by type of troponin. Despite the fact that troponin essays have longed being considered equivalent in the acute coronary syndrome setting [34], emergency evidence supports that hs-cTnT and hs-cTnI outcome prognostic value might differ in other scenarios [35,36]. To assess if such a difference exists in this case, we will group studies by the declared essay type (i.e. hs-cTnI, hs-cTnT, hs-cTnC).
- Subgroup by type of association. In order to assess whether heterogeneity may be due to the type of analysis employed to derive a dichotomous effect size, a subgroup analysis will be performed segregating studies reporting dose-response data (i.e. regressions handling hs-cTn as a continuous dependent variable) versus category/quantile-based data (i.e. contingency tables and/or regressions handling hs-cTn as a dichotomous dependent variable). This subgrouping is not applicable for the SMD model.
Furthermore, to assess whether missing adjustment for relevant confounders might be impacting the heterogeneity and pooled effect size of the mortality adjusted model: We will undertake a sensitivity analysis, meta-analyzing only effect measures of studies adjusting for (at least) the 3 predefined core confounders.
To avoid p-value overreliance, the credibility of the differential effect of statistically significant Subgroup analysis will be critically appraised with the ICEMAN questionnaire [37].
If we suspect a small study effect in any model, we will run a sensitivity analysis with a fixed effect model.
2.8. Publication bias
For assessment of possible small-study effect, funnel plots and Eggers test will be run for any given model that incorporates at least 10 different studies, by plotting the logarithm of the obtained odds ratios against their standard error [19,38–40]We will visually inspect the plots for signs of publication bias (namely, asymmetry
3. Discussion and conclusions
The prognostic significance of troponin for sepsis outcomes is not yet settled. Although previous works have consistently shown a significant association between troponin and mortality, there is a need for more research shedding light on the prognostic value of such biomarker for the most common current clinical setting in the ICU. This study will provide further evidence about the role of hs-cTn assays in predicting mortality in septic patients; potentially helping to guide further research and yielding valuable information for patient assessment.
The systematic review outline in this protocol is not exempt of possible difficulties and limitations: most notably, the failure to declare critical data, and/or the sensitivity of the troponin essays utilized by primary studies might result excluding a substantial amount of the eligible pool; another problem with troponin essays is the multiplicity of normality cutoffs, something that could end up in a synthesis of cohorts categorizing patients at different risk levels [19]; moreover, as with any meta-analysis of adjusted estimates, we expect to synthetize cohorts that considered different sets of confounders, which can introduce substantial heterogeneity in the final estimate [19].
We will evaluate the certainty of evidence for each investigated outcome with the ‘Grading of Recommendations Assessment, Development and Evaluation’ (GRADE) system guideline appropriate for synthesis of prognostic factors evidence [41,42]. A “Summary of findings’ table will communicate with standardized statements the certainty of this evidence
Supporting information
S1 Checklist. PRISMA-P (Preferred Reporting Items for Systematic review and Meta-Analysis Protocols) 2015 checklist: Recommended items to address in a systematic review protocol*.
https://doi.org/10.1371/journal.pone.0301948.s001
(DOC)
References
- 1. Mayr FB, Yende S, Angus DC. Epidemiology of severe sepsis. Virulence. 2014 Jan 1;5(1):4–11. pmid:24335434
- 2. Martin GS, Mannino DM, Eaton S, Moss M. The epidemiology of sepsis in the United States from 1979 through 2000. N Engl J Med. 2003 Apr 17;348(16):1546–54. pmid:12700374
- 3. Kumar G, Kumar N, Taneja A, Kaleekal T, Tarima S, McGinley E, et al. Nationwide trends of severe sepsis in the 21st century (2000–2007). Chest. 2011 Nov;140(5):1223–31. pmid:21852297
- 4. Singer M, Deutschman CS, Seymour CW, Shankar-Hari M, Annane D, Bauer M, et al. The Third International Consensus Definitions for Sepsis and Septic Shock (Sepsis-3). JAMA. 2016 Feb 23;315(8):801–10. pmid:26903338
- 5. Eriksson J, Eriksson M, Brattström O, Hellgren E, Friman O, Gidlöf A, et al. Comparison of the sepsis-2 and sepsis-3 definitions in severely injured trauma patients. Journal of Critical Care. 2019 Dec 1; 54:125–9. pmid:31442842
- 6. Font MD, Thyagarajan B, Khanna AK. Sepsis and Septic Shock—Basics of diagnosis, pathophysiology and clinical decision making. Med Clin North Am. 2020 Jul;104(4):573–85. pmid:32505253
- 7. Ehrman RR, Sullivan AN, Favot MJ, Sherwin RL, Reynolds CA, Abidov A, et al. Pathophysiology, echocardiographic evaluation, biomarker findings, and prognostic implications of septic cardiomyopathy: a review of the literature. Crit Care. 2018 Dec;22(1):112. pmid:29724231
- 8. L’Heureux M, Sternberg M, Brath L, Turlington J, Kashiouris MG. Sepsis-Induced Cardiomyopathy: a Comprehensive Review. Curr Cardiol Rep. 2020;22(5):35. pmid:32377972
- 9. Sato R, Nasu M. A review of sepsis-induced cardiomyopathy. Journal of Intensive Care. 2015 Nov 11;3(1):48. pmid:26566443
- 10. Collet JP, Thiele H, Barbato E, Barthélémy O, Bauersachs J, Bhatt DL, et al. 2020 ESC Guidelines for the management of acute coronary syndromes in patients presenting without persistent ST-segment elevation. European Heart Journal. 2021 Apr 7;42(14):1289–367.
- 11. Twerenbold R, Jaffe A, Reichlin T, Reiter M, Mueller C. High-sensitive troponin T measurements: what do we gain and what are the challenges? European Heart Journal. 2012 Mar 1;33(5):579–86. pmid:22267244
- 12. The Septic Heart: Current Understanding of Molecular Mechanisms and Clinical Implications. Chest. 2019 Feb 1;155(2):427–37.
- 13. Sheyin O, Davies O, Duan W, Perez X. The prognostic significance of troponin elevation in patients with sepsis: a meta-analysis. Heart Lung. 2015;44(1):75–81. pmid:25453390
- 14. Bessière F, Khenifer S, Dubourg J, Durieu I, Lega JC. Prognostic value of troponins in sepsis: a meta-analysis. Intensive Care Med. 2013 Jul;39(7):1181–9. pmid:23595497
- 15. Zheng P, Wang X, Guo T, Gao W, Huang Q, Yang J, et al. Cardiac troponin as a prognosticator of mortality in patients with sepsis: A systematic review and meta-analysis. Immun Inflamm Dis. 2023 Sep;11(9): e1014. pmid:37773717
- 16. Røsjø H, Varpula M, Hagve TA, Karlsson S, Ruokonen E, Pettilä V, et al. Circulating high sensitivity troponin T in severe sepsis and septic shock: distribution, associated factors, and relation to outcome. Intensive Care Med. 2011 Jan;37(1):77–85. pmid:20938765
- 17. Page MJ, McKenzie JE, Bossuyt PM, Boutron I, Hoffmann TC, Mulrow CD, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ. 2021 Mar 29; 372:n71. pmid:33782057
- 18. Moons KGM, de Groot JAH, Bouwmeester W, Vergouwe Y, Mallett S, Altman DG, et al. Critical Appraisal and Data Extraction for Systematic Reviews of Prediction Modelling Studies: The CHARMS Checklist. PLoS Med. 2014 Oct 14;11(10): e1001744. pmid:25314315
- 19. Riley RD, Moons KGM, Snell KIE, Ensor J, Hooft L, Altman DG, et al. A guide to systematic review and meta-analysis of prognostic factor studies. BMJ. 2019 Jan 30; 364: k4597. pmid:30700442
- 20. Ouzzani M, Hammady H, Fedorowicz Z, Elmagarmid A. Rayyan—a web and mobile app for systematic reviews. Systematic Reviews. 2016 Dec 5;5(1):210. pmid:27919275
- 21. American College of Chest Physicians/Society of Critical Care Medicine Consensus Conference: Definitions for sepsis and organ failure and guidelines for the use of innovative therapies in sepsis. Critical Care Medicine. 1992 Jun;20(6):864.
- 22. Levy MM, Fink MP, Marshall JC, Abraham E, Angus D, Cook D, et al. 2001 SCCM/ESICM/ACCP/ATS/SIS International Sepsis Definitions Conference. Crit Care Med. 2003 Apr;31(4):1250–6. pmid:12682500
- 23. Bekkering GE, Harris RJ, Thomas S, Mayer AMB, Beynon R, Ness AR, et al. How Much of the Data Published in Observational Studies of the Association between Diet and Prostate or Bladder Cancer Is Usable for Meta-Analysis? American Journal of Epidemiology. 2008 May 1;167(9):1017–26. pmid:18403406
- 24.
Higgins JPT, Cochrane Collaboration, editors. Cochrane handbook for systematic reviews of interventions. Second edition. Hoboken, NJ: Wiley-Blackwell; 2020. (Cochrane book series).
- 25. Rohatgi A. Webplotdigitizer: Version 4.6 [Internet]. 2022. Available from: https://automeris.io/WebPlotDigitizer.
- 26. Grooten WJA, Tseli E, Äng BO, Boersma K, Stålnacke BM, Gerdle B, et al. Elaborating on the assessment of the risk of bias in prognostic studies in pain rehabilitation using QUIPS—aspects of interrater agreement. Diagnostic and Prognostic Research. 2019 Mar 7;3(1):5. pmid:31093575
- 27. Hamling J, Lee P, Weitkunat R, Ambühl M. Facilitating meta‐analyses by deriving relative effect and precision estimates for alternative comparisons from a set of estimates presented by exposure level or disease category. Statistics in Medicine. 2008 Mar 30;27(7):954–70. pmid:17676579
- 28. Xu KZ, Xu P, Li JJ, Zuo AF, Wang SB, Han F. Predictors and nomogram of in-hospital mortality in sepsis-induced myocardial injury: a retrospective cohort study. BMC Anesthesiol. 2023 Jul 7;23(1):230. pmid:37420185
- 29. Vallabhajosyula S, Sakhuja A, Geske JB, Kumar M, Poterucha JT, Kashyap R, et al. Role of Admission Troponin-T and Serial Troponin-T Testing in Predicting Outcomes in Severe Sepsis and Septic Shock. J Am Heart Assoc. 2017 Sep 9;6(9): e005930. pmid:28889100
- 30. Vallabhajosyula S, Sakhuja A, Geske JB, Kumar M, Kashyap R, Kashani K, et al. Clinical profile and outcomes of acute cardiorenal syndrome type-5 in sepsis: An eight-year cohort study. PLoS One. 2018 Jan 9;13(1): e0190965. pmid:29315332
- 31. Symons MJ, Moore DT. Hazard rate ratio and prospective epidemiological studies. J Clin Epidemiol. 2002 Sep;55(9):893–9. pmid:12393077
- 32. Viechtbauer W. Conducting meta-analyses in R with the metafor package. Journal of Statistical Software. 2010;36(3):1–48.
- 33. Guyatt G, Zhao Y, Mayer M, Briel M, Mustafa R, Izcovich A, et al. GRADE guidance 36: updates to GRADE’s approach to addressing inconsistency. Journal of Clinical Epidemiology. 2023 Jun 1; 158:70–83. pmid:36898507
- 34. van der Linden N, Wildi K, Twerenbold R, Pickering JW, Than M, Cullen L, et al. Combining High-Sensitivity Cardiac Troponin I and Cardiac Troponin T in the Early Diagnosis of Acute Myocardial Infarction. Circulation. 2018 Sep 4;138(10):989–99. pmid:29691270
- 35.
American College of Cardiology [Internet]. [cited 2024 Jan 21]. Utility of High-Sensitivity Troponin T and I: Are They the Same? Available from: https://www.acc.org/Latest-in-Cardiology/Articles/2019/11/21/07/26/http%3a%2f%2fwww.acc.org%2fLatest-in-Cardiology%2fArticles%2f2019%2f11%2f21%2f07%2f26%2fUtility-of-High-Sensitivity-Troponin-T-and-I.
- 36. Willeit P, Welsh P, Evans JDW, Tschiderer L, Boachie C, Jukema JW, et al. High-Sensitivity Cardiac Troponin Concentration and Risk of First-Ever Cardiovascular Outcomes in 154,052 Participants. J Am Coll Cardiol. 2017 Aug 1;70(5):558–68.
- 37. Schandelmaier S, Briel M, Varadhan R, Schmid CH, Devasenapathy N, Hayward RA, et al. Development of the Instrument to assess the Credibility of Effect Modification Analyses (ICEMAN) in randomized controlled trials and meta-analyses. CMAJ. 2020 Aug 10;192(32): E901–6. pmid:32778601
- 38. Sterne JA, Egger M. Funnel plots for detecting bias in meta-analysis: guidelines on choice of axis. J Clin Epidemiol. 2001 Oct;54(10):1046–55. pmid:11576817
- 39. Sterne JAC, Sutton AJ, Ioannidis JPA, Terrin N, Jones DR, Lau J, et al. Recommendations for examining and interpreting funnel plot asymmetry in meta-analyses of randomised controlled trials. BMJ. 2011 Jul 22;343(jul22 1): d4002–d4002. pmid:21784880
- 40. Egger M, Davey Smith G, Schneider M, Minder C. Bias in meta-analysis detected by a simple, graphical test. BMJ. 1997 Sep 13;315(7109):629–34. pmid:9310563
- 41. Foroutan F, Guyatt G, Zuk V, Vandvik PO, Alba AC, Mustafa R, et al. GRADE Guidelines 28: Use of GRADE for the assessment of evidence about prognostic factors: rating certainty in identification of groups of patients with different absolute risks. Journal of Clinical Epidemiology. 2020 May 1; 121:62–70. pmid:31982539
- 42. Iorio A, Spencer FA, Falavigna M, Alba C, Lang E, Burnand B, et al. Use of GRADE for assessment of evidence about prognosis: rating confidence in estimates of event rates in broad categories of patients. BMJ. 2015 Mar 16;350:h870. pmid:25775931