Introduction

Aortic stenosis became the most common primary valve disease in the western world mainly due to the progressive aging of the population. Symptomatic aortic stenosis has a high mortality rate and carries a significant healthcare burden [1,2,3].

It is already established that the degenerative disease affecting the aortic valve is the result of an active process where inflammation and calcification, associated with cardiovascular risk factors, play a key role [1, 4,5,6].

The positron emission tomography-computed tomography (PET-CT) with 18F-sodium fluoride (18F-NaF) has long been used to evaluate bone metastases [7]. Recently this tracer has been applied, in clinical research, to characterize atherosclerotic plaques and aortic valve disease [8,9,10,11]. It is specific for microcalcification through ion changes between the hydroxyl group and the hydroxyapatite chrystals. It is known that calcification is a common pass way in aortic valve degeneration and atherosclerotic plaque progression [12]. Vascular calcification occurs, mainly, through osteogenic conversion of mesenchymal tissue into calcified bone. Regulators of bone formation are upregulated by activation of immune cells recruited to sites of chronic inflammation. Lipid oxidation promoted by cytoquines and free radicals intervene in the activation of osteogenic regulatory genes leading to the differentiation of mesenchymal progenitor cells. The mesenchymal-derived osteoblasts secrete a collagen proteoglycan matrix with the ability of binding calcium and phosphate salts, forming hydroxyapatite minerals [12].

Our aim was to characterize through PET-CT the uptake of 18F-NaF by the aortic valve in patients with risk factors for cardiovascular disease and evaluate its relation with cardiovascular risk assessed by the ASCVD risk calculator [13].

Material and methods

Population

Hypertensive patients, age over 40 years old, were recruited from the outpatient clinic of a tertiary university hospital. The inclusion was made upon the acceptance to participate in this prospective and exploratory study from May 1, 2014 to June 1, 2015.

Inclusion criteria

Besides hypertension patients were included if they had other primary risk factors, chronic kidney disease with a glomerular filtration rate below 60 ml/min [according Modification of Diet in Renal Disease equation—(MDRD)], diabetes mellitus or a markedly abnormal single risk factor.

Exclusion criteria

Patients under 40 years old and childbearing age women were excluded. Patients with cardiovascular disease (previous cardiac disease, cerebrovascular disease or peripheral vascular disease), with symptoms suggestive of disease (angina, heart failure symptoms, neurological complaints or claudication), aortic stenosis, severe renal impairment (GFR < 30 ml/min), moderate to severe hepatic failure (Child–Pugh class B or C), chronic inflammatory or neoplasic diseases were also excluded.

Upon inclusion, a baseline assessment of CVR was performed and for that purpose a blood sample was drawn to evaluate the levels of total cholesterol, LDL and HDL cholesterol, triglycerides, fasting blood glucose, creatinine, hemoglobin A1c (HbA1c) and C-reactive protein (CRP).

All patients underwent an 18F-NaF PET-CT to assess the uptake of the tracer in the aortic valve in the same day of clinical, echocardiographic and biochemical evaluation. Echocardiography was performed to rule out aortic valve disease, namely, aortic stenosis.

All the procedures were done with the approval of the ethics committee of the Faculty of Medicine of the University of Coimbra, in accordance with the Declaration of Helsinki and upon the written informed consent of each participant.

PET-CT

A 16-slice PET-CT system was used (Gemini GXL Philips) and images were acquired 60 min after intravenous administration of 185 MBq of 18F-NaF (5 mCi). The enrolled patients underwent cardiac PET-CT imaging. After the performance of an attenuation correction CT scan (non-enhanced 120 kV and 50 mA), PET images were acquired. Iterative reconstruction was obtained in multiple phases. A matrix of 144 × 144 was applied for reconstruction of transversal PET slices. Voxel size was of 4 × 4 × 4 mm and the spatial resolution of 8 mm. Focal uptake of the tracer was determined by PET-CT fusion images. Aortic valve uptake was evaluated in a 3-dimensional multiplanar mode and 4 mm slices, in horizontal transaxial planes, were considered top to bottom of the aortic valve for the establishment of circular regions of interest (ROI) around the valve. Semi-quantitative analysis was accomplished using the standardized uptake values (SUV) estimated for each slice and for the whole valve (Fig. 1). These values were corrected for blood-pool activity by subtraction—corrected uptake per lesion (CUL) as previously described by the authors [11]. Blood-pool SUV was estimated as the mean of five ROI in the mid lumen of superior vena cava. From the CT images, Agatston calcium score of the aortic valve was also evaluated.

Fig. 1
figure 1

Fusion 18F-NaF-PET-CT images, depicting aortic valve 18F-NaF uptake

Full size image

Statistical analysis

Continuous data was tested for normality with the Shapiro-Wilks test and was represented by their mean, median, and interquartile range. Student’s t test and the Mann–Whitney test were applied to compare continuous variables as appropriate. Correlations between normally distributed data were performed with Pearson correlation and presented as r2 values. Spearman correlation was used for nonparametric data.

Data analysis was performed by StatView 5.0.1, version for Macintosh and Windows, SAS Institute.

Results

All the included 25 patients were followed by the outpatient clinic due to hypertension. Their mean age was of 63.9 ± 8.6, minimum of 48 and maximum of 80 years old. Figure 2 shows the distribution by gender and cardiovascular risk factors. Clinical evaluation and biochemical parameters are expressed in Table 1. In general, this is a high cardiovascular risk population where 80% have diabetes and 72% hyperlipidemia. The median BMI was also of 29.56. Total cholesterol, LDL cholesterol, fasting blood glucose and triglycerides were also above the normal values in almost all patients.

Fig. 2
figure 2

Distribution of the population according to gender and cardiovascular risk factors

Full size image
Table 1 Population clinical, biochemical and ASCVD assessment
Full size table

Echocardiographic assessment of the aortic valve showed no significant stenosis as the maximum velocity was inferior to 2 m/s in all patients [14].

The calcium score of the valve was of 0 in 11 patients (44%), with a mean of 44.40 ± 165.7, median 6.00 and IQR 29.75. 18F-NaF uptake was estimated through the corrected uptake by lesion (CUL)and was 0.5 ± 0.1, median 0.52 and IQR 0.15.

The uptake of the tracer and the calcium score of the valve were correlated with ASCVD considering the risk as a continuous variable and as a categorical variable with two defined risk groups: less than the 50th percentile or greater or equal than the 50th percentile. The results are presented in Fig. 3 and Table 2. Simple regression analysis didn’t show a significant relationship between the assessed variables but a trend is noticed between cardiovascular risk and 18F-NaF uptake by the aortic valve (Fig. 3). When the population was divided in two groups according to the 50th percentile of the ASCVD, the 18F-NaF uptake by the valve was significantly higher in higher risk patients, while the calcium score didn’t show a significant difference between the groups (Table 2). We also analyze the uptake of the tracer regarding the presence of calcium and no significant differences were found. Those without any calcium had an uptake of 0.47 ± 0.18 (median 0.48; IQR 0.22) comparing with 0.51 ± 0.11 (median 0.52; IQR 0.12) of those with some degree of valve calcification by the calcium score (p = 0.54).

Fig. 3
figure 3

Simple regression graphics: a ASCVD and AoVCUL, b ASCVD and AoVCaSc, c AoVCaSc and AoVCUL (ASCVD risk score atherosclerotic cardiovascular disease risk score, AoVCUL aortic valve corrected uptake per lesion, AoVCaSc aortic valve calcium score)

Full size image
Table 2 Comparison between AoVCUL and the AoVCaSc with the ASCVD considering distribution of patients according the 50th percentile
Full size table

Discussion

The natural history of aortic valve degeneration comprises a long asymptomatic period during which occurs fibrotic thickening of the aortic leaflets with progression to calcification which is the most common cause of aortic stenosis affecting 0.4% of the general population. The asymptomatic period is variable and is usually associated with a risk of sudden death of less than 1% [1].

Aortic valve leaflets comprise three layers: covering the aortic side, a collagen rich fibrosa, a middle proteoglican-rich spongiosa and covering the ventricular side, the elastin-rich ventricularis. The leaflets are covered by valvular endothelial cells (VEC) and fibroblast-like cells, the valvular interstitial cells (VIC), are spread between the described layers. The remodeling of the valve begins by the action of several factors such age, biochemical, genetic and mechanical risk factors. Activation of VEC, inflammation, angiogenesis and VIC transformation by osteogenic differentiation contribute to the degeneration process [15,16,17].

The association between aortic valve disease and cardiovascular risk factors suggests that aortic stenosis has an atherosclerotic-like progression. Nowadays it is known that a more complex process is involved with specific-cell signaling pathways regulating valvular calcification [15,16,17].

Fluorine-18 fluorodeoxyglucose(18F-FDG) and 18F-NaF uptake have been used to evaluate inflammation and microcalcification in aortic valve disease and the results are promising since they target key mediators for disease progression [18]. In what respects 18F-FDG it seems that its uptake was higher in patients with mild to moderate aortic stenosis in comparison with controls or with those with severe disease [19]. Abdelbaky A et al. concluded that 18F-FDG uptake was a marker of disease progression as those patients with higher rates of valve calcification were those with higher 18F-FDG early uptake [20]. In observational retrospective studies, on oncologic and rheumatologic patients, 18F-NaF uptake by the aortic valve was higher in patients with aortic calcification and stenosis [21].

Dweck et al.in a prospective study published in 2012 showed that 18F-FDG has a higher uptake in aortic stenosis patients than in controls increasing discreetly with the severity of the disease [6]. In comparison the 18F-NaF uptake was higher in patients with sclerosis and aortic stenosis than in controls and increases with disease severity. In this study 35% and 91% of patients with aortic stenosis had 18F-FDG and 18F-NaF uptake, respectively [6]. They also conclude through the follow-up of a subset of these patients that the uptake of 18F-NaF correlates with calcium deposition seeming to predict disease progression [8]. This follow-up study comprises a group of patients in whom aortic valve surgery was performed and on those an association between 18F-NaF signal and markers of calcification (alkaline phosphatase and osteocalcin) was found [8].

Giving the above results we aimed to assess, as far as we know by the first time, the uptake of 18F-NaF by the aortic valve of patients with high cardiovascular risk. Aortic stenosis was excluded in all of them. We used the CUL instead of TBR in the quantification of aortic valve uptake of the tracer considering the results already published [11, 22].

It was found that 18F-NaF uptake and though microcalcification were related with ASCVD risk leading to the hypothesis that sodium fluoride could be a biomarker of disease onset. Calcification estimated by calcium score was not related with ASCVD risk and several patients had a score of 0 in spite of their cardiovascular risk.

These results seemed to support the association between risk factors and aortic valve microcalcification [23, 24]. In fact, before stenosis, aortic valve sclerosis was found to be related with 18F-NaF uptake [6]. In this study, 18F-NaF uptake was observed in patients with no detected calcification and no stenosis of the aortic valve.

Previous studies targeting the relation between cardiovascular risk factors and aortic disease had no positive results [25,26,27] and this was probably linked with the presence of advanced disease characterized by calcification and stenosis. In fact, the presence of advanced degeneration of the valve imposes mechanical stress and osteogenic differentiation that further increase disease progression [15, 16]. Microcalcification, according to our results, seemed to be associated with cardiovascular risk assessed by the ASCVD risk calculator and could be an early marker of valve degeneration.

Limitations

The number of patients studied was small, and so, our results and conclusions should be considered hypothesis generating needed to be validated in a larger population. In order to evaluate aortic disease progression, a long follow-up period is required.

The ideal dose of F18-NaF for this kind of study remains to be established, quantification of sodium fluoride uptake is still controversial and the used outcome measure could influence conclusions.

New knowledge gained

In patients with no significant valvular disease, microcalcification, in particularly, from aortic valve could be measured with 18F-NaF PET-CT.

18F-NaF uptake of aortic valve was correlated with cardiovascular risk measured with ASCVD score.

No significant correlation was found between aortic valve calcium score and 18F-NaF uptake or with ASCVD score.

Conclusion

As far as we know this study was the first attempt to characterize the aortic valve in high cardiovascular risk patients. In spite of the small sample size it shows the existence of a link between aortic uptake of sodium fluoride and cardiovascular risk. Further studies with larger populations are required to confirm the described results.