Author + information
- Published online May 15, 2019.
- Stefania Rosmini, MD, PhD,
- Heerajnarain Bulluck, MBBS, PhD,
- Amna Abdel-Gadir, MBBS,
- Thomas A. Treibel, MBBS, PhD,
- Veronica Culotta, MD,
- Richard Thompson, PhD,
- Stefan K. Piechnik, PhD, MSCEE,
- Peter Kellman, PhD,
- Charlotte Manisty, PhD and
- James C. Moon, MD∗ ()
- ↵∗Barts Heart Centre, St. Bartholomew’s Hospital, West Smithfield, London EC1A 7BE, United Kingdom
Native myocardial T1 and extracellular volume (ECV) by cardiovascular magnetic resonance imaging are key biomarkers for assessing several heart muscle diseases. Their measurement, however, is influenced by blood T1, either due to myocardial vascular blood volume or via blood pool partial voluming. Different factors, including hematocrit, pH, temperature, sex, and oxygen pressure, are known to affect blood T1 (1,2) but the influence of blood biological constitutents such as iron, lipids, or albumin is unknown.
Our goal was to assess this topic. With appropriate ethical approval and written informed consent, we prospectively recruited 77 healthy volunteers free of cardiovascular disease or diabetes, with normal 12-lead electrocardiography and normal cardiovascular magnetic resonance imaging. T1 mapping was performed at 1.5-T using all 3 primary approaches to the technique: MOLLI (MOdified Look-Locker Inversion recovery) (3), ShMOLLI (Shortened MOdified Look-Locker Inversion recovery) (4), and SASHA (Saturation recovery Single-sHot Acquisition) (5). Just before the scan, a blood sample was drawn and hemoglobin, hematocrit, N-terminal pro–B-type natriuretic peptide, creatinine, estimated glomerular filtration rate, iron, total iron-binding capacity, ferritin, albumin, glycosylated hemoglobin, and a nonfasting lipid profile (including triglycerides, total cholesterol, low-density lipoprotein cholesterol, and high-density lipoprotein [HDL] cholesterol) were tested. T1 values were measured in blood and myocardium before and after administration of extracellular magnetic resonance contrast medium (3). Univariable and multivariable linear regression analyses were performed to identify biochemical factors affecting blood T1.
The population mean age was 49 ± 14 years (range: 20 to 76 years). Thirty-eight (49%) of the 77 patients were male (mean age: male subjects 50 ± 14 years, female subjects 49 ± 15 years). For MOLLI, ShMOLLI, and SASHA, respectively, myocardial T1s were 1,027 ± 38 ms, 959 ± 31 ms, and 1,144 ± 48 ms; blood T1s were 1,638 ± 78 ms, 1,543 ± 77 ms, and 1,587 ± 103 ms. As expected, as hematocrit rose, blood T1 fell for all sequences: ShMOLLI, r2 = 0.530, p < 0.0001; MOLLI, r2 = 0.452, p < 0.0001; and SASHA, r2 = 0.375, p < 0.0001.
However, univariable analysis (here ShMOLLI–other sequence concordant) showed additional correlations with iron bound to transferrin (r2 = 0.405; p < 0.0001), HDL cholesterol (r2 = 0.229; p < 0.0001), creatinine (r2 = 0.115; p < 0.002), ferritin (r2 = 0.101; p = 0.003), triglycerides (r2 = 0.097; p = 0.003), and low-density lipoprotein cholesterol (r2 = 0.031; p = 0.069). Albumin, total cholesterol, N-terminal pro–B-type natriuretic peptide, estimated glomerular filtration rate, and glycosylated hemoglobin were not associated with blood T1 (by any sequences, log-transformed where needed).
Multivariable analysis showed that blood T1 is largely determined by just 3 variables, in descending order of effect: hematocrit, iron, and HDL cholesterol regardless of sequences (MOLLI: cumulative r2 = 0.778; ShMOLLI: r2 = 0.772; SASHA: r2 = 0.567; all p < 0.0001), with SASHA correlation being slightly lower than that of MOLLI and ShMOLLI (Figure 1).
Measured blood T1 was associated with measured myocardial T1 (correlation r2 = 0.266, p = 0.01 by MOLLI; r2 = 0.374, p = 0.001 by ShMOLLI; and r2 = 0.578, p < 0.0001 by SASHA). ECV is similarly associated with blood T1 (r2 = 0.392, p = 0.001 by MOLLI; r2 = 0.464 by ShMOLLI; and r2 = 0.457 by SASHA with p < 0.0001, respectively).
To summarize, blood myocardial T1 is primarily determined by blood biology, with 3 parameters (hematocrit, iron, and HDL cholesterol) of the ones investigated here explaining almost 80% measured variability in health. Anemia, iron deficiency, and a high HDL cholesterol, all more common in women, increase blood T1 and appear to explain this typical female, longer blood T1. Native blood T1 is associated with myocardial T1 and ECV. This may be from the inclusion of blood signal in measured myocardial T1 but also potentially from biological processes (e.g., compensatory vasodilation in anemia). Further research is required to better understand this finding.
Please note: This work was supported by Borse di studio SIC e MSD Italia-Merck Sharp & Dohme, the Rosetrees Trust, GlaxoSmithKline, and the National Institute for Health Research Oxford Biomedical Research. The research was conducted at The Heart Hospital, London (which has now merged with Barts Heart Centre). The authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- 2019 American College of Cardiology Foundation