Author + information
- Alberto Aimo, MD,
- Giuseppe Vergaro, MD, PhD and
- Michele Emdin, MD, PhD∗ ()
- ↵∗Address for correspondence:
Scuola Superiore Sant’Anna and Fondazione Toscana Gabriele Monasterio, Via Giuseppe Moruzzi 1, Pisa 56124, Italy
We read with interest the meta-analysis by Evans et al. (1), reporting that high-sensitivity (hs) troponins help predict new-onset heart failure (HF), which included as many as 67,063 patients with 4,165 incident HF events. In the accompanying editorial comment, Drs. Jaffe and Miller focus on the large differences in the analytical properties of hs-troponin assays, which they consider the main reason of the observed between-study heterogeneity (2). We suggest the population was heterogeneous for other respects. First, 10 studies of 16 evaluated general populations (i.e., patients not selected based on their risk profile), whereas the other 6 included patients with conditions predisposing to HF: diabetes mellitus (2 studies), hypertension (1 study), stable cardiovascular disease (1 study), or chronic kidney disease (1 study) (1). Further, 3 of the 6 studies were also nested in randomized clinical trials (1), which notoriously represent a different setting than real-world studies. It may also be considered that a HF diagnosis represented an exclusion criterion, but only 2 studies explicitly excluded patients with left ventricular ejection fraction <50% or as low as 40% (1). Heterogeneity was observed with regard to the endpoint definition: “HF hospitalization alone,” “fatal or nonfatal HF diagnoses and events,” or “all nonfatal HF diagnoses and events” (1).
Following the recent completion of an individual patient data meta-analysis on hs-troponin T and prognosis in chronic HF (3), we required participating authors provide individual patient datum from the original studies. This would allow to consider hs-troponin concentration as a continuous variable, to directly assess the distribution of hs-troponin levels, and to search for hs-troponin cutoffs that can be considered for HF screening, instead of comparing “participants in the top third with those in the bottom third of baseline values” based on a statistical elaboration and specific assumptions (1). In addition, the prognostic value of hs-troponins could be evaluated on top of a clearly defined set of variables, instead of having to use a rather vague formula such as “demographic factors, conventional cardiovascular disease risk factors, and/or biomarkers other than natriuretic peptides […] or natriuretic peptide levels” (1).
In conclusion, Evans et al. (1) should be congratulated for their original work; nevertheless, the intrinsic limitations of their report, common in meta-analyses, should be recognized, especially in a period when meta-analyses are dismissed as “medical fake news” (4). International collaborations aiming to share individual patient data may give more insight on the potential of hs-troponin assays to predict future development of HF, either in the general population or in specific categories of patients at high risk for HF.
Please note: The authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- 2018 American College of Cardiology Foundation
- Evans J.D.W.,
- Dobbin S.J.H.,
- Pettit S.J.,
- Di Angelantonio E.,
- Willeit P.
- Jaffe A.S.,
- Miller W.L.
- Aimo A.,
- Januzzi J.L.,
- Vergaro G.,
- et al.
- Packer M.