Influence of hydration status on cardiovascular magnetic resonance myocardial T1 and T2 relaxation time assessment: an intraindividual study in healthy subjects

Myocardial native T1 and T2 mapping allow for a non-invasive quantification of myocardial tissue alterations across a broad range of myocardial diseases in patients eligible for cardiovascular magnetic resonance (CMR) imaging [1]. As myocardial mapping is very sensitive for the detection of myocardial tissue abnormalities like myocardial edema or fibrosis, current guidelines (e.g. for inflammatory cardiomyopathies) already recommend the implementation of mapping techniques into clinical routine [2]. However, it is important to know that mapping techniques can be prone to confounders such as heart rate and magnetic field inhomogeneities [1, 3, 4]. Apart from these variations caused by confounders, there are also subtle differences in myocardial T1 and T2 relaxation times that are related to gender or age [5]. Moreover, myocardial iron content can drastically alter myocardial relaxation times [4]. Although it is well known that myocardial T1 and T2 relaxation are sensitive to pathological increase of myocardial water content (e.g. myocardial edema) [6‐8], the influence of physiological hydration changes as a possible confounder of relaxation time assessment has not been fully evaluated yet. Although previous experimental studies suggest that changes in CMR relaxation times occur with non-pathological alterations in myocardial water content [9], no actual in vivo data is available about the relationship between changes in myocardial water content and concordant myocardial native T1 and T2 relaxation time measurements. Thus, the purpose of this study was to evaluate whether the change in myocardial water content due to dehydration and hydration might alter the correct assessment of myocardial relaxation times in healthy subjects.

The institutional review board approved this prospective study and all subjects gave written informed consent. The study population consisted of healthy subjects without known cardiovascular disease and without cardiovascular disease risk factors. All subjects had an unremarkable past medical history of cardiovascular disease. Electrocardiographic results prior to CMR were unremarkable. All control participants had normal CMR results without structural abnormalities or wall motion abnormalities. To assess possible influences of hydration status on native T1 and T2 relaxation time measurements, three separate CMR scans were performed in each participant: The baseline scan (normal scan) was performed during the morning hours. Participants were allowed to have normal breakfast at home, but coffee intake was restricted. Also, they were told not to ingest significant amounts of water prior to the baseline scan. After 12 h of fasting (no access to food or water) another scan was performed (dehydration scan). The last scan was performed after adequate hydration (hydration scan). After the end of fasting each participant was offered a standardized small cheese sandwich. For hydration, the participants were instructed to drink 1.5 l of still water within 15 min. Drinking was supervised by a study team member. The hydration scan was started 15 min after fluid intake to allow water absorption. For all participants the scans were performed at 1 day. Blood pressure measurements and laboratory markers were obtained directly prior to every scan. Blood samples were promptly transported to the central laboratory and were immediately analyzed. A visualized description of the study protocol is presented in Fig. 1. For the assessment of inter-study reproducibility another group of healthy subjects was studied. Inter-study reproducibility was measured from two separate examinations that were acquired within two consecutive CMR examinations during the same day.

A total of 12 healthy subjects were enrolled (5 male). Mean age was 25.8 ± 3.2 years (men: 28.0 ± 3.2; range 25–32; women: 24.3 ± 2.1; range 24–26; P = 0.063). Mean body mass index was 22.9 ± 3.5 kg/m² (mean height: 175.2 ± 8.9 cm, mean weight: 70.6 ± 12.8 kg). All participants completed the study protocol. Structural CMR results were normal for all participants.

In this prospective study, we evaluated whether differences in hydration status in healthy subjects could influence the assessment of myocardial native T1 and T2 relaxation times. The main finding of this study is that myocardial native T1 and T2 values were significantly reduced in dehydration, indicating an influence of hydration status on relaxation time assessment. Therefore, hydration status might be a confounder of native T1 and T2 mapping.

Dehydration (deficit of body water) is a general physiological condition that can have profound effects on human health [12]. Mild dehydration often only leads to slight impairment in physical or mental performance and can be easily corrected [13]. In our study we observed significantly increased levels of hematocrit and red as well as white blood cell count indicating measurable effect of dehydration. Also, left ventricular end-diastolic volume index was reduced due to the restrictions in fluid intake. Chronic dehydration can have serious effects on human health and might be associated with or might promote common public health disorders [14]. Especially, older adults are susceptible to dehydration due to age-related pathophysiological changes. When hospitalized, dehydrated older patients (aged ≥65 years) are 6 times more likely to die in hospital than patients with a normal level of hydration. However, in a clinical setting dehydration is often underdiagnosed. Uncomplicated dehydration (e.g. due to water deprivation) results in a significant loss of tissue water compared to normal hydrated controls, especially in the muscle and the skin [15].

Although different tissues with similar water content have been shown to have significantly different native T1 and T2 relaxation time [16], to some extent T1 and T2 relaxation time are a function of absolute tissue water content and correlate strongly with tissue water content [9, 17]. Investigating the same tissue (e.g. intervertebral disc anulus fibrosus and nucleus pulposus), correlations of relaxation times are mostly influenced by differences in tissue water content [17]. In a tissue analysis animal study from 1984, Brown et al. investigated T1 and T2 relaxation times of normal, volume overloaded, and dehydrated rabbit myocardium and found that native T1 and T2 values significantly correlated with percent tissue water content [9]. Compared to normal rabbits they found significantly reduced native T1 and T2 relaxation time in rabbit heart tissues, which were dehydrated by water restriction and furosemide-induced diuresis [9]. Our results indicate that also in healthy humans, differences in native T1 and T2 relaxation time can be measured depending on hydration status. Particularly in a dehydrated state, significant alterations in relaxation times compared to baseline measurements were observed. During dehydration myocardial relaxations times were reduced compared to baseline. Water loss due to water deprivation or inadequate water intake typically leads to hypertonic dehydration, which can be also caused by sweat loss (e.g. during exercise and fever) and osmotic diarrhea [12]. In hypertonic dehydration, water loss exceeds sodium loss, which is characterized by an osmotic shift of water from the intracellular to the extracellular compartment. The relatively lower amount of water in the intracellular compartment might have influenced myocardial native T1 and T2 relaxation time measurements observed in our study.

Although differences in myocardial native T1 relaxation times after dehydration compared to baseline were rather small (mean difference: 19 ± 19 ms), the mean difference was higher than the 95% of agreement of inter- and intra-observer and the inter-study reproducibility measurements, indicating a potential important effect when evaluating relaxation times in a clinical setting. However, the influence of dehydration in a clinical context has not been evaluated yet. According to a consensus statement about the clinical use of CMR mapping, local reference ranges for myocardial mapping should be obtained from healthy subjects and only values below or above two standard deviations should be considered as pathological [1]. The standard deviations for T1 mapping in this study ranged from 27 ms to 29 ms. In this regard, dehydration might not lead to pathological results in a predominant number of patients in a sample, but may possibly influence the result of the relaxation measurement in individual cases. Interestingly, during the hydration CMR we were not able to show increased myocardial relaxation times compared to baseline, possibly because the chosen intervention (drinking 1.5 l of water) was not sufficient to show a hyperhydration effect in the healthy subjects. In another CMR study using T1 mapping, Graham-Brown et al. investigated interstudy reproducibility in 10 hemodialysis patients and found T1 relaxation times to be unaffected from fluid status [18]. However, the study results cannot be generalized as no systematic scans (e.g. before and after dialysis) were performed in this study and the reported differences in Δ weight (as surrogate for the altered fluid status) were close to zero for the most patients. In our study, we showed decreased relaxation times healthy controls using strict study protocol and therefore random effects on fluid status, as in the study from Graham-Brown et al. [18], can be excluded. On the other hand, native T1 and T2 times are prolonged with increasing water content of tissue and the presence of myocardial edema, especially in myocardial infarction, myocarditis or other acute cardiomyopathies [6, 7, 19‐21]. In this regard it is important to know that alterations in fluid status may in theory affect the assessment of relaxation times leading to confounding results. Hydration status might be an additional confounder the clinician should have in mind when interpreting results, especially in follow-up studies and in cases of subtle myocardial disease or fibrosis [22‐25].

Hydration status may influence myocardial native T1 and T2 relaxation time. In our study, dehydration led to significant alterations in relaxation times and therefore may influence precise or repeatable assessment of native T1 and T2 relaxation times. Hydration status should be recognized as new potential confounder of T1 and T2 relaxation time assessment in clinical routine. Future studies should assess the influence of dehydration on CMR relaxation time measurements in a clinical setting.