Consecutive patients at different stages of HH [diagnosed based on clinical characteristics, abnormal iron turnover parameters, and the presence of HFE gene mutations (9)] were prospectively enrolled in the study from October 2015 to November 2018. The exclusion criteria were age < 18 years, history of any cardiac diagnosis (apart from hypertension), features of heart damage, and left ventricular ejection fraction (LVEF) < 50%. The control group comprised healthy age- and sex-matched volunteers. All participants underwent echocardiography, and detailed medical histories including duration of HH and administered treatments with laboratory parameters were obtained. The study protocol was approved by the Local Ethics Committee at the Medical University of Gdańsk (NBBN/452/2016), and written informed consent was obtained from all the participants.

The patients were divided into three groups according to the time of diagnosis and initiation of treatment to investigate the influence of HH duration on the heart. The “recently diagnosed” HH group consisted of patients diagnosed with HH for less than 6 months and had no more than one venesection. The “medium” group consisted of patients who had HH diagnosed between 6 and 24 months (2 years). Finally, the “long-lasting” group consisted of patients diagnosed with HH for more than 2 years. All the parameters in these groups were analyzed and compared according to the age distribution obtained during the statistical analyses to investigate the impact of age.

All patients underwent echocardiography at the time of enrollment. Patients were examined in the left lateral decubitus position using a GE VIVID E95 ultrasound system (GE Ultrasound, Horten, Norway) equipped with a phased-array transducer (M5S). Standard echocardiographic parameters were obtained according to the guidelines of the American Society of Echocardiography (ASE) and the European Association of Cardiovascular Imaging (EACVI) recommendations (12, 13). Data acquisition was obtained from the parasternal long- and short-axis views and the three standard apical views. Three consecutive cardiac cycles were recorded during quiet respiration for each view. Grayscale recordings were optimized for LV evaluation at a rate of 50–80 frames/s, and only patients with these parameters were included in the subsequent analyses. All echocardiograms were digitally stored, and further offline analysis was performed using a commercial EchoPAC workstation (v204, GE Healthcare, Horten, Norway).

Continuous data were presented as the medians (25th–75th percentiles), whereas categorical data were expressed as proportions. We performed the Shapiro-Wilk test to determine the normal distribution of our data. Most of the analyzed parameters did not have normal data distributions, even after logarithmic data transformation; therefore, we selected appropriate statistical analysis methods based on non-parametric tests. The significance of differences between the patients with HH and the control group was assessed using the Wilcoxon test (between patients with HH and controls and between groups of patients divided by disease duration and age distribution based on the obtained quartiles). Comparisons between all groups were performed using the Kruskal–Wallis test for continuous variables (with Dunn's post-hoc test for the multiple comparisons with Bonferroni adjustment to determine the significantly different pairs of groups and control for the overall error rate when conducting multiple comparisons) or by the chi-squared test or Fisher's test for categorical variables. Statistical significance was set at p-values <0.05 were considered significant. Statistical analyses were performed using R version 4.2.1. (R Core Team, Vienna, Austria).

Fifty-eight consecutive patients diagnosed with HH between 1 month to 20 years were enrolled in the study. Forty-one patients had the C282Y/C282Y mutation, 12 had the C282Y/H63D mutation, four had the H63D/H63D mutation, and one had the C282Y/WT mutation. The median patient age was 45 years (range: 31–57 years). Table 1 shows the demographic data, medical history, and laboratory results of all the patients with HH. Twenty-nine age- and sex-matched healthy volunteers comprised the control group.

Table 2 presents a comparison of the echocardiographic examination between patients with HH and healthy individuals. Some parameters related to diastolic function [LA size (LADs, LAV index), IVS, PW, RWT, and LVM index] were significantly worse in patients with HH than in healthy individuals (Figure 1). The LVEF was within the normal range; however, it was significantly lower in patients with HH. There were no significant differences in the LV and the right ventricle sizes. We found differences in many 2D STE parameters, including those related to diastolic function (LV untwisting rate and velocity), and in many related to systolic function (LV apex rotation, global twist and torsion, and peak systolic longitudinal strain; Figure 2).

Table 3 shows the demographic data, medical histories, and laboratory results of patients with HH at the different stages of the disease. There was an increasing trend in the age and frequency of hypertension in the analyzed HH groups, and the rate of glucose disturbances increased noticeably in the long-lasting HH group. Iron levels were stable between the groups with reduced ferritin concentrations in medium and long-lasting HH patients. TSAT and transaminases levels were similar between groups.

Table 4 presents the comparisons of the standard and 2D STE echocardiographic parameters between HH patients in terms of disease stage. Significant differences in LA size were observed between the controls and each HH group but not between the patients with HH from each group (Figure 3). Differences in LV wall thickness, RWT, and LVM index were the most prominent between healthy volunteers and patients with long-lasting HH (Figure 3). Differences in LVEF were the most evident between the healthy volunteers, the recently diagnosed group, and the medium HH group (Table 4). Regarding the 2D STE parameters, the medium group had significantly better rotation indices than the recently diagnosed patients, with no significant differences compared to the long-lasting HH group. Regarding peak systolic longitudinal strain, significantly worse values were noticed in all the HH groups compared to the controls, with no significant changes between the HH groups: −19.3% [−20.5–−17.3] in the recently diagnosed group, −17.5% [−18.2–−15.9] in the medium group, and −18.0% [−19.6–−15.6] in the long-lasting HH group, contrary to −21.0% [−22.0–−19.3] in the control group (Figure 4).

Figures 5, 6 present three HH patients from each group with graphical demonstration of twist and peak systolic longitudinal strains.

We divided patients with HH into four groups according to age quartiles based on age distribution to investigate the influence of the age on heart changes: the quartile from minimum to Q25 included patients from 18 to 31 years; Q25–Q50 included patients from 32 to 47 years; Q50–Q75 included patients from 48 to 56 years old; and the fourth quartile (Q75 to maximum) included patients from 57 to 77 years. Table 5 shows the demographic data, medical history, and laboratory results of the patients with HH stratified by age. Among the patients with HH, the incidence of hypertension and diabetes increased with age. The rate of glucose disturbances increased in patients aged 32–47 years. Iron levels were stable between the groups, with reduced ferritin concentrations among patients aged 32–47 years and 57–77 years. The levels of TSAT and transaminases were comparable in both.

Table 6 compares the standard and 2D STE echocardiographic parameters of patients with HH stratified by age. Some standard parameters related to diastolic function (LA size, IVS, RWT, LVM index, Em, and E/Em) changed above the normal range most prominently in the third decade of life (Figure 7). In contrast, some STE parameters related to systolic function (LV global twist and torsion) increased significantly from the fourth decade of life (Figure 8).

Table 7 presents the correlations between iron turnover parameters and measured echocardiographic parameters. Some correlations were observed between the time from diagnosis and the number of venesections with echocardiographic parameters. The correlations were significant; however, they were relatively weak (the r-values were lower than 0.7).

Standard echocardiography revealed differences in diastolic parameters in the enrolled patients with HH compared with healthy controls. Several authors have documented diastolic dysfunction in patients with HH (14–16). The cited studies' techniques were based on the standard two-dimensional echocardiography, including M-mode (16) and TDI techniques (14, 15). For example, Candell-Riera et al., showed that patients with idiopathic hemochromatosis presented significantly higher LADs index (22.2 ± 3.7 mm/m²) and LVM index (150 ± 56.2 g/m²) compared to the control group (respectively 19.2 ± 1.7 mm/m² and 91.7 ± 14.6 g/m², respectively) (16). Similarly, Palka et al. reveled deviations in the LADs and the LVM index in their analysis of patients with HH, with 83% being C282Y homozygotes (15). The authors also observed significantly higher values of LVM index (101 ± 26 g/m²) and LADs in patients with HH (3.81 ± 0.74 cm) compared to healthy controls (77 ± 12 g/m² for LVM index and 3.27 ± 0.39 cm for LADs, respectively) (15). Our results are in line with this study: we found significantly higher LADs [42.0 (36.0–43.0) mm/m²] and LAV index values [31.0 (23.0–36.5) ml/m²], and LVM index values [78.0 (58.0–96.0) g/m²] in patients with HH compared to healthy volunteers (37.0 [34.0–39.0] mm/m2 for LADs, 21.5 [19.0–27.1] ml/m2 for LAV index, and 66.0 [53.0–72.0] g/m2 for LVM index, respectively) (Figure 1). Interestingly, Davidsen et al. did not show such differences in the LADs and LVM index values between patients with HH and healthy controls, possibly because all of their patients were regularly treated with venesections (14).

Furthermore, we compared patients with HH at the different stages of the disease. To the best of our knowledge, this is the first such analysis of patients with HH. We revealed worse diastolic parameters measured by standard echocardiography in patients with HH compared to controls, with the most noticeable differences (in morphological and functional parameters) detected between the healthy and long-lasting HH group. The latter had worse LV wall thickness, RWT, LVM index, and Em indices, probably due to the disease's length and the patient's age (Table 4, Figure 3). Hypertension and glucose disturbances, which occur earlier in patients with HH than in the overall population, could have an additional impact on the left ventricular diastolic function (17).

The application of 2D STE in our study allowed for a more precise description of systolic dysfunction. In one of our previous studies, the patients with recently diagnosed HH had worse rotation and strain parameters measured by 2D STE than healthy controls, despite the lack of differences in standard echocardiographic parameters (Figure 4) (10). We confirmed this observation in the presented study of patients with HH at the different stages of the disease.

2D STE may allow for the complex analysis of heart mechanics. A decrease in the untwisting velocity represents a predominance of diastolic dysfunction, and significant decreases in rotation, rotation rate, twist, torsion, and peak systolic longitudinal strain are critical predictors of systolic dysfunction (18–20). Only a few studies (10, 11, 21–23) have used speckle-tracking analysis in the context of iron-overload symptoms. However, these studies involved patients with beta-thalassemia major, an extreme model of systemic iron overload (24). In our study, the recently diagnosed HH group was characterized by the worst parameters of the LV rotation, the twisting and untwisting indices, whereas the percentages of hypertension and glucose disturbances were the lowest in that group compared to the other HH groups (Table 4; Figure 4), which may exclude the possible influence of these comorbidities on 2D STE parameters. The aforementioned rotation indices improved significantly in the medium and long-lasting groups of patients, possibly due to HH treatment (Table 4; Figure 4). This postulation aligns with previous data in the literature (11, 25). For example, Byrne et al. showed that intensification of venesection therapy may significantly improve radial strains following a 1-year course of phlebotomies (11). In one of our clinical case presentations, we observed similar changes in the rotation, twist, and torsion LV parameters after 6-month therapy with venesections in our patient with recently diagnosed HH (25). However, we did not inestigate the direct influence of treatment but compared the patients at the different HH stages. Our results make it unclear why the Long-lasting HH patients did not present further improvement in rotation parameters. That could be explained in two ways: the initial treatment by venesections could only normalize rotation parameters in HH patients without further improvement, and the aging of the enrolled patients could implicate our results (26). However, these explanations must be verified in a further prospective study with the appropriate follow-up observations.

The peak systolic longitudinal strain values in this study were significantly lower in patients with HH at all stages of the disease (Figure 4). This parameter did not improve with treatment initiation and continuation, which suggests early irreversible changes in the hearts of patients with HH starting at the early stages of the disease. We obtained similar results in our previously published case report, where peak systolic longitudinal strain did not improve within the treatment (25). This finding may have a substantial clinical impact since an impaired peak longitudinal strain value is a well-established prognostic factor for cardiac complications (27–29). However, further research is necessary to verify whether the differences in the peak systolic longitudinal strain between the healthy volunteers and patients with HH at different disease stages are clinically relevant. The LVEF was within the normal range in our patients, despite statistically significant differences between the patients with HH and controls. This finding is consistent with previously documented data (16), indicating that LVEF may not be the best parameter for assessing systolic damages in the patients with HH.

Iron overload may accelerate the aging of the cardiovascular system, including the heart, through oxidative stress, which results in appropriate changes in diastolic and systolic parameters. These changes can be detected by standard echocardiography (increase in the LV width and LVM index, LA enlargement, or changes in mitral flow parameters) or by STE (decrease in rotation, twist, and torsion values of the LV) (10, 15–17).

In the presented study, the echocardiographic features of myocardial walls' thickening in patients with HH appeared more rapidly than in healthy idividuals. For instance, Ganau et al. evaluated the effect of age on concentric remodeling in healthy participants and showed that RWT values increased with age and constituted 0.32 ± 0.044 among patients younger than < 41 years, 0.34 ± 0.05 for those aged 41–64 years, and 0.37 ± 0.050 for those > 64 years (30). Our patients with HH achieved similar RWT values faster than healthy individuals in the cited study (Figure 7). In our study, the LVM index was 76 (62–91) g/m² in the fourth decade, whereas Kaku et al. reported a lower maximum value of the LVM index in the older (the eighth decade of life) healthy volunteers (31). Another important parameter that increases with age is the LA size. Nikitin et al. showed that the LAD was 3.51 ± 0.51 cm in patients aged 20–39 years, 3.99 ± 0.54 cm in those aged 40–59 years, 4.12 ± 0.43 cm in those aged 60–79 years, and 4.19 ± 0.51 in those aged > 80 years (32). In our analysis, the LADs increased slightly faster (Figure 7). Similarly, the LAV index increased faster in our patients (Figure 7) than in healthy individuals, as reported by Nikitin et al. (32). Furthermore, we showed increased E/Em values in the older patients compared to the younger patients; however, we did not find any significant differences from the general population (31).

Regarding LVEF, as an index of systolic function, Kaku et al. showed that it does not significantly change with age (31). Similarly, we did not observe any changes in LVEF in our patients with HH with age; however, LVEF was slightly lower than that of the general population (31). STE may allow for a better understanding of age-related systolic function changes than a standard echocardiography with LVEF measurments. According to Kuznetsova et al., the absolute values of the peak systolic longitudinal strain significantly decreased in a healthy population with age: 24.2 ± 3.4%, in individuals aged < 40 years, 23.5 ± 3.19% in those aged 40–59, and 22.3 ± 3.62% in those aged > 60 years (33). Among our patients with HH, peak systolic longitudinal strain was at a lower level: −17.8 (−19.1–16.2) % in patients with HH aged 18–31 years, and was without significant changes in the older patients (among patients aged 32–47: −19.8 [−21.0–17.1] %, aged 48–56: 18.3 [−19.2–16.9] %, aged 57–77: 19.6 [−20.1–17.0]%; Figure 8). Kaku et al. reported that the LV global twist and torsion values increased with age (34). According to that study, the maximum LV twist and torsion values were reached in the seventh decade of life (12.7 ± 4.8° for LV twist and 1.75 ± 0.66°/cm for torsion). In our analysis, similar LV twist values were present in patients with HH at a younger age (18–31 years old) and constituted 11.1 (9.4–17.8)° for twist and 1.5 (1.2–2.0)°/cm for torsion (Figure 8). Increasing the rotation parameters within the age range may result in a decrease in the longitudinal strain of the LV.

Our results showed that thickening of the myocardium and dilatation of the LA occurred faster in patients with HH than in the general population. Similarly, the deterioration in systolic function detected using STE was more prominent in patients with HH than in the general population. This could suggest a harmful effect of the disease on the heart in addition to the aging of the patients; however, this hypothesis needs to be verified.

The presented analysis showed that the correlations between the iron turnover parameters and echocardiographic indices were relatively weak. These findings align with data from the literature (23, 35, 36) and our previous results, including cardiac magnetic resonance assessments (10, 37). There is probably no “direct” relationship between the levels of iron turnover parameters and myocardial function, and myocardial iron overload is not the only mechanism involved in the development of HH cardiomyopathy (1). The precise mechanism of heart involvement in HH is yet to be entirely understood. The excessive capacity of serum transferrin to bind iron, which is typical in patients with HH, results in uncontrolled iron entry into cardiomyocytes, which may increase their susceptibility to oxidative cell stress (3). Cardiomyocytes contain many mitochondria, causing the myocardium to be more vulnerable to oxidative stress damage. Free iron ions that damage the mitochondrial and nuclear DNA may activate fibroblast proliferation and differentiation into myofibroblasts, which are responsible for heart fibrosis (38). Oxidative stress results in the peroxidation of cell membranes, particularly those in the mitochondria, which reduces the amount of adenosine triphosphate generated during oxidative phosphorylation. Oxidative stress impairs heart muscle relaxation and delays contraction by inhibiting SERCA2 enzyme activity and increasing the cytoplasmic concentration of calcium ions in cardiomyocytes. Therefore, amlodipine, a Ca channel blocker, may be promising as an agent to potentially reduce oxidative stress during iron overload (39).

Magnetic resonance imaging is the gold standard for measuring iron overload (2). Previous studies presented changes in the heart documented on cardiac magnetic resonance (CMR) in patients with beta-thalassemia, most of whom met the current criteria for iron overload in terms of myocardial T2 * < 20 ms (40–42). HH is not the same iron overload model as beta-thalassemia, which has been documented in our previous study (37). In all the studied patients with HH, myocardial T2*, T1, and T2* did not fulfill the clinically severe iron overload (20 ms) threshold, which confirms the hypothesis that in modern patients with HH not simple iron storage in the heart could play an essential role in the heart damage process. Therefore, in the present study, we assessed the role of echocardiography, particularly STE, in identifying differences in patients with HH at various disease stages and ages, rather than calculating iron overload.