Background Cusp calcification is the main mechanism leading to bioprosthetic heart valve (BPV) failure. Recent studies suggest that BPV calcification is an active rather than passive process probably modulated by several mechanisms including lipid-mediated inflammation and dysfunctional phosphocalcic metabolism.
Objective To identify the clinical and metabolic determinants of BPV calcification assessed by multidetector CT (MDCT).
Methods and results Presence of BPV calcification was assessed by MDCT in 194 patients who had undergone aortic valve replacement. A calcification score was individually calculated and expressed in mm3. Patients also underwent a clinical evaluation, a Doppler echocardiographic exam, and a plasma lipid and phosphocalcic profile. 46 patients (24%) had BPV calcification (cusp calcification score >0 mm3). After adjustment for age, gender, and time interval since BPV implantation, increased calcium–phosphorus product (OR 1.11, 95% CI 1.01 to 1.23 per 1 unit; p=0.02) and the presence of prosthesis-patient mismatch (OR 3.67, 95% CI 1.25 to 10.6; p=0.01) were the strongest independent factors associated with BPV calcification. Calcium supplement intake, age and female gender were independently associated with increased calcium–phosphorus product.
Conclusions This study suggests that higher calcium–phosphorus product and prosthesis–patient mismatch promote BPV calcification. Furthermore, this study reports that calcium supplements, which are extensively prescribed in elderly patients, are independently associated with higher calcium–phosphorus product.
Statistics from Altmetric.com
The main advantage of bioprosthetic valves (BPV), compared to mechanical valves, is their lower thrombogenicity which obviates the need for lifelong anticoagulation. Nonetheless, their use is still limited by their durability with a rate of reoperation for structural failure of 20–30% 15 years after implantation.1 Calcification plays a major role in the failure of BPVs. The underlying pathologic process is cuspal mineralisation leading to cusp stiffening and progressive valve stenosis and/or cusp tearing leading to valve regurgitation. Recent studies suggest that BPV calcification is a complex process possibly modulated by several mechanisms, including lipid-mediated inflammation, immune response, and a dysfunctional phosphocalcic metabolism.2 ,3
Multidetector CT (MDCT) has been shown to be a reproducible and accurate quantitative method to assess native aortic valve calcification in patients with aortic stenosis.4 The aim of this prospective study was to identify the clinical and metabolic determinants of BPV calcification as measured by MDCT.
Between June 2008 and June 2010, we prospectively recruited 203 patients who underwent an isolated aortic valve replacement with a bioprosthesis at the Quebec Heart & Lung Institute (QHLI) at least 3 years previously. All patients recruited underwent Doppler echocardiography and 194 patients underwent MDCT. MDCT could not be performed or provided inadequate images in nine patients (4%) and these patients had similar baseline characteristics compared to patients with MDCT data available. All patients gave written inform consent approved by the institutional review board of the QHLI. The population characteristics and methods of this study have been previously reported.5
Patients also had a clinical examination and a complete plasma glycaemic, lipid and phosphocalcic profile generally the same day they underwent MDCT and Doppler echocardiographic exams (or if not, within 10 days). Patients were excluded if they were younger than 21 years, were pregnant or lactating women, and/or if they had congestive heart failure (New York Heart Association functional class III or IV). Echocardiographic exclusion criteria were as follows: (1) Presence of >mild paravalvular regurgitation; (2) significant concomitant mitral valve disease; (3) subvalvular flow acceleration precluding measurement of BPV valve effective orifice area (EOA); (4) LV systolic dysfunction defined by an LVEF <50%.
Clinical and operative data
Clinical data including age, gender, anthropometric measures, cardiovascular risk factors, comorbidities, and medications were collected. Operative data including model and size of bioprosthetic valves were also recorded. Prosthesis-patient mismatch (PPM) was defined as not clinically significant (ie, mild or no PPM) if the indexed projected EOA was >0.85 cm2/m2, moderate if it was >0.65 cm2/m2 and ≤0.85 cm2/m2, and severe if it was ≤0.65 cm2/m2.
The echocardiograms were all performed and reviewed by the same cardiologist (HM) using high-quality commercially available ultrasound systems (IE33, Philips, Andover, Massachusetts, USA and Vivid 7, GE Healthcare, Little Chalfont, UK). All patients underwent a complete comprehensive echocardiographic examination with assessment of BPV haemodynamic profile and chamber size and function. Prosthetic valve regurgitation was detected by colour Doppler echocardiography and the origin of the jet was visualised in several views to differentiate periprosthetic from transprosthetic valve regurgitation. LVEF was calculated using the modified biplane Simpson's method.
The CT was performed using a 64 slice helical scanner (Somatom Definition, Siemens AG Medical Solution, Germany). The entire heart was assessed by 3 mm thick axial slices with a pitch of 0.35 and B35f kernel during held inspiration. No iterative reconstruction was used. Acquisitions were obtained with a tube potential at 120 kV and a tube current-time product at 80 mAS. Operators blinded to patient data performed all MDCT examinations and analyses. Off-line analysis of the cardiac images was conducted using an automated, computerised software program (Aquarius, TeraRecon Inc, San Mateo, California, USA). Bioprosthesis calcification was measured by the volumetric method which identifies calcium within the BPV cusps as areas of at least two contiguous pixels with a density of 130 HU or more as previously described.6 The total volume of calcifications was calculated and a calcification score was then determined for each patient and expressed in mm3 (figure 1).
Particular attention was paid to avoid artefacts due to prosthesis sewing rings, and to distinguish calcifications located in the region of the BPV cusps from those located in the region of the prosthesis sewing ring and aortic annulus.
All blood samples were obtained after an overnight fast and plasma was isolated and immediately processed by the laboratory for the measurement of total calcium, inorganic phosphate and creatinine by standard photometric Roche Diagnostics methods on Modular analyser (Roche, Diagnostics GmbH, Manheim, Germany). The glycaemic and lipid profile was also assessed as previously described by Mahjoub et al.5 The glomerular filtration rate (GFR) was calculated using the Cockcroft-Gault and the modified diet in renal disease formula. Moderate insufficiency was defined as GFR ≥30 and <60 mL/min/1.73 m2 and severe insufficiency was defined as GFR <30 mL/min/1.73 m2. Both serum 25-OH vitamin D and serum intact parathyroid hormone (iPTH) were measured by electrochemiluminescent analysis on a Modular Analytics E170 (Roche Diagnostics). Vitamin D status was defined as normal if the serum 25-OH vitamin D concentration was ≥50 nmol/L, as vitamin D insufficiency if it was ≥30 nmol/L and <50 nmol/L, and as vitamin D deficiency if it was <30 nmol/L. Hyperparathyroidism was defined as iPTH serum concentrations ≥65 ng/L and secondary hyperparathyroidism as iPTH serum concentrations ≥65 ng/L with concomitant low 25-OH vitamin D serum values (≤50 nmol/L) and normal total calcium values (≥2.55 nmol/L).
Patients were separated into two groups depending on presence or absence of BPV calcification (calcification volume > or =0 mm3, respectively). Categorical data were expressed as numbers and percentages and compared with one-way analysis of variance, χ2 test or Fisher's exact test as appropriate. Continuous data were expressed as mean±SD and compared using the unpaired Student t test.
Correlations between variables were determined using Pearson (continuous variables) or Spearman (categorical variables) methods. A multiple logistic regression analysis was used to identify the factors independently associated with BPV calcification. Factors independently associated with calcium–phosphorus product were identified with the use of stepwise multiple linear regression. Age as well as variables with a value of p<0.1 on univariable analysis were entered into the multivariate models. A value of p<0.05 was considered statistically significant.
Baseline characteristics of the study population
One hundred and ninety-four patients were included in this study. Mean age of the study population was 67±8 years and a mean time interval since BPV implantation was 7.9±3 years (range 3.0–18.6 years); 69% were male, 47% had a history of coronary artery disease (CAD), and 44% had a moderate or severe renal insufficiency (table 1). Forty-six patients (24%) had evidence of BPV calcification, and in this group the median value of BPV calcium volume was 1.58 mm3 and the interquartile range (IQR) was 4.22 mm3. Time interval since BPV implantation was significantly longer (9.2±4.2 vs 7.5±3.1 years; p=0.005) in patients with BPV calcification compared to patients with no BPV calcification. There was no significant difference between the two groups with regard to other cardiovascular risk factors and medications. As for operative data, there was no difference between both groups regarding the type (porcine vs bovine, or stentless vs stented) and size of BPV. However, there was a trend for higher prevalence of moderate or severe PPM (indexed projected EOA≤0.85 cm2/m2) in the BPV calcification group compared to the no BPV calcification group (17% vs 7%; p=0.06).
Patients with BPV calcification had higher mean transprosthetic gradients (19±7 vs 13±7 mm Hg; p<0.0001) and a higher prevalence of significant transprosthetic valve regurgitation (21% vs 2%; p=0.04) than patients with no BPV calcification. Moreover, when assessing the progression of mean transprosthetic gradients between the echocardiographic examination done within the first postoperative year and the echocardiographic examination done in the context of this prospective study, we found that patients with BPV calcification also had faster progression rate of mean gradient since 1-year postoperative echocardiogram (absolute increase: +7±7 vs +2±5 mm Hg; p<0.0001 and annualised increase +0.98 ±1.2 vs +0.32 mm Hg/year; p=0.001).
Patients with BPV calcification had significantly higher calcium–phosphorus product (2.55±0.36 vs 2.41±0.39 mmol2/L2; p=0.04) (table 2). There were no significant difference in iPTH and 25-OH vitamin D serum concentrations between the two groups. Furthermore, prevalence of vitamin D insufficiency and of hyperparathyroidism were, respectively, 11% and 32% in the BPV calcification group and 14% and 29% in the no BPV calcification group (p=0.7; p=0.6, respectively).
Factors associated with bioprosthetic valve calcification
On univariable logistic regression analysis, time interval since BPV implantation (p=0.005), diabetes (p=0.03), calcium–phosphorus product (p=0.04), and PPM (p=0.06) were associated with BPV calcification. On multivariable analysis, the independent determinants of BPV calcification were time interval since BPV implantation (p=0.003), calcium–phosphorus product (p=0.02), and PPM (p=0.01) (table 3). When gender was forced into the model, time interval since BPV implantation (OR 1.16, 95% CI 1.05 to 1.29; p=0.003), calcium–phosphorus product (OR 1.11, 95% CI 1.01 to 1.23 per 0.1 increment; p=0.03), and PPM (OR 3.63, 95% CI 1.2 to 10.8; p=0.02) remained independent determinants of BPV calcification. With further adjustment for diabetes, results were similar: time interval since BPV implantation (OR 1.15, 95% CI 1.04 to 1.28; p=0.003), calcium–phosphorus product (OR 1.12, 95% CI 1.01 to 1.24 per 0.1 increment; p=0.04), and PPM (OR 3.20, 95% CI 1.07 to 9.4; p=0.05).
Factors associated with higher calcium–phosphorus product
As calcium–phosphorus product was the strongest metabolic predictor of BPV calcification, we sought to determine the clinical and metabolic factors associated with increased calcium–phosphorus product in our study population who had mostly normal calcium–phosphorus product values (mean 2.45±0.39 mmol2/L2; range 1.45 to 3.64 mmol2/L2). Higher calcium–phosphorus product was positively associated with age (p=0.004), female gender (p<0.0001), calcium supplement use (p<0.0001), vitamin D supplement use (p<0.0001), and bisphosphonate use (p=0.0003). On multivariable analysis, age (p=0.01), female gender (p<0.0001), and calcium supplement use (p=0.002) were independently associated with higher calcium–phosphorus product (table 4).
In a previous study of the same cohort of patients,5 we reported that 42 patients had evidence of BPV haemodynamic dysfunction at echocardiography. In the present study, 46 patients had BPV calcification at MDCT. However, these two groups of patients are different as only 19 patients have both BPV calcification and dysfunction. To this effect, Schoen and Levy7 reported that BPV failure may occur in the absence of leaflet calcification and, vice versa, calcification may be present without any associated dysfunction. BPV calcification often precedes valve dysfunction in the natural history of BPV degeneration. In the present study, we elected to focus on the determinants of BPV calcification given that this is an early marker of BPV structural degeneration. The main findings of this study are: (1) Higher calcium–phosphorus product and prosthesis–patient mismatch are independent determinants of BPV calcification; and (2) calcium supplements are independently associated with higher calcium–phosphorus product in this elderly population with normal calcium–phosphorus product values.
Impact of phosphocalcic metabolism on BPV calcification
For years, calcification of BPVs was seen as a purely passive process triggered by the chemical fixative treatment of bovine or porcine tissue and enhanced by mechanical stress.7 ,8 Nonetheless, recent analyses of explanted BPVs and valve allografts reveal that inflammatory cells and bone regulatory proteins are present in valve leaflets.7 ,9 These findings suggest that BPV calcification is not solely a random degenerative process but may also involve active osteogenic mechanisms. Recent studies documented that high phosphate serum concentrations can trigger the osteoblastic differentiation of vascular smooth muscle cells and can stimulate the production of several osteogenic factors such as Cbfa-1 and osteopontin.10 Moreover, in the large Cardiovascular Health Study, higher phosphate concentrations within the normal range were associated with a higher prevalence of valvular calcification.11 These findings underline the potential important contribution of phosphate in vascular and valvular calcification, and may explain, at least in part, its independent association between phosphataemia and cardiovascular morbidity.12 The present study is the first to demonstrate that elevated calcium–phosphorus product is a strong independent determinant of BPV calcification, which may in turn influence cardiovascular morbidity. However, the vast majority of the patients, including those with BPV calcification, had values of calcium–phosphorus product within the normal range. This finding suggests that even a subtle alteration of the phosphocalcic homeostasis could play a role in accelerated BPV calcification. It is also possible that elevated calcium–phosphorus product is simply a marker for other factors involved in the BPV mineralisation.
Emerging evidence in the literature suggests a link between ectopic calcification and altered bone mineralisation metabolism.13–15 The measure of bone mineral density was not available in the present cohort. Given that only 18 patients in this study took bisphosphonates (ie, the treatment generally prescribed for osteoporosis), we can assume that the prevalence of severe osteoporosis in this cohort was probably low. Nonetheless, it is likely that in such a population a relatively large proportion of the patients had mild to moderate osteoporosis.
The paradoxical association between bone demineralisation—that is, osteopenia or osteoporosis—and ectopic calcification is supported by the results of several studies.16 ,17 It has also been suggested that ectopic calcification is a possible explanation of the increased cardiovascular mortality associated with a lower bone mineral density.
The mechanisms underlying this paradoxical association remains largely unknown. Intuitively, excessive bone resorption increases the availability of calcium and phosphate (ie, calcium-phosphorus product) in the blood and can thus lead to a passive deposition of calcium–phosphorus crystals in vascular or valvular tissue. Other mechanisms have been proposed to explain this paradoxical association, including the proposal that osteoporosis and ectopic calcification may share common causative factors such as dyslipidaemia and inflammation. To this effect, it has been suggested that oxidised low density lipoprotein (LDL) may have opposite effects in vessels or valves versus bones inducing calcification in vascular/valvular tissues and demineralisation in bone tissues.18
Impact of calcium supplements on phosphocalcic product
Another important finding of this study is that calcium supplementation was independently associated with higher calcium–phosphorus product, which was found to be a strong determinant of BPV calcification. These findings raise the concern that calcium supplements may accelerate BPV calcification. Calcium supplements have traditionally been considered a safe alternative to meet calcium requirements. However, recent studies have raised concerns about potential adverse cardiovascular effects of calcium supplementation,19 ,20 while other studies have reported no significant impact on cardiovascular outcomes.21 ,22 Several pots-hoc analyses of randomised clinical trials showed a significant effect of this supplementation on cardiovascular events.23 ,24 It has been hypothesised that the negative impact of calcium supplementation on cardiovascular outcomes is related to vascular calcification leading to arterial stiffness, and also to valvular calcification and ensuing stenosis. This adverse effect could be related to the fact that calcium supplements, unlike dietary calcium, could induce an abrupt change in calcium blood values that may favour its deposit in vascular or valvular tissues. Nonetheless, other recent studies have not found any significant association between calcium supplementation and vascular calcification25 or progression of aortic valve calcification.26 The results of the present study further add to previous concerns about the potential negative effect of calcium supplementation on ectopic calcification. Given the very large number of elderly people who currently take calcium supplements, further studies are urgently needed to determine the actual impact of calcium supplementation on vascular and valvular calcification.
Impact of prosthesis–patient mismatch on BPV calcification
The present study also reveals that PPM is a strong predictor of BPV calcification. These results confirm and expand those of Flameng et al reporting that PPM is a strong independent predictor of BPV stenosis due to structural valve degeneration.27 PPM occurs when the EOA of a normally functioning prosthesis is too small in relation to the patient's body size and thus to cardiac output requirements. This results in abnormally high transprosthetic gradients and disturbance of haemodynamic flow patterns, which may, in turn, increase the mechanical stress on BPV cusps and thus accelerate BPV calcification.7 ,8 To this effect, several histological studies of BPVs explanted for structural valve degeneration revealed that calcifications are located at the base and/or near the commissure in sites of intense mechanical deformations generated by motion such as the points of leaflet flexion.2 ,3 ,8 Our results suggest that PPM may contribute to the calcifying process of BPVs and therefore underline the need to avoid PPM in patients undergoing aortic valve replacement (AVR) with a BPV.
The present study is a prospective cross-sectional study that reported the prevalence of BPV calcification at the time of the study, which limits our ability to assess the value of calcium-phosphorus product to predict future development of BPV calcification. Additional studies with longitudinal follow-up of blood biomarkers of phosphocalcic metabolism and of BPV calcification are necessary to confirm the usefulness of the calcium–phosphorus product to predict BPV mineralisation. Furthermore, the presence of a significant association between calcium–phosphorus product and BPV calcification does not necessarily imply causality.
We did not find a direct association between BPV calcification and the use of calcium supplement. This may be due to the relatively small sample size with only 44 patients taking calcium supplements. The same limitation also applies to other medications which could potentially influence BPV calcification such as bisphosphonates.
This study suggests that higher calcium–phosphorus product and prosthesis–patient mismatch promote BPV calcification. The calcium–phosphorus product was nonetheless within the normal range in all patients included in this study, which suggests that even subtle alterations in phosphocalcic metabolism may promote BPV calcification. Moreover, we found that the use of calcium supplements is independently related to higher calcium–phosphorus product, which raises the hypothesis that this widely prescribed supplementation may accelerate BPV calcification. These findings further support the need for the prevention of PPM at the time of AVR and for the realisation of large scale studies to elucidate the impact of calcium supplementation in elderly populations with BPVs.
What is known on this subject?
Mineralisation of valve cups is the main mechanism leading to structural degeneration of bioprosthetic heart valves (BPVs). Mulidectector CT provides a unique opportunity to assess BPV calcification.
What might this study add?
This study reveals that higher calcium–phosphorus product and prosthesis–patient mismatch are the main independent determinants of BPV calcification measured by MDCT. Furthermore, calcium supplements which are extensively prescribed in elderly patients are independently associated with higher calcium–phosphorus product.
How might this impact on clinical practice?
These findings support the fact that the prevention of prosthesis–patient mismatch at the time of aortic valve replacement could influence BPV calcification and also raise the concern that calcium supplements may induce dysfunctional phosphocalcic metabolism that could in turn induce BPV calcification.
The authors would like to thank Jacinthe Aubé, Martine Poulin and Dominique Labrèche for data collection and technical assistance.
Contributors HM performed data collection, echo and statistical analyses and wrote the manuscript with the help of PP and PM. ÉL performed computed tomography analyses. J-PD supervised blood plasma analyses and collection of biological data. All authors reviewed the manuscript.
Funding This study was funded by a research grants (MOP #86666 and MOP#114893) from the Canadian Institutes of Health Research (CIHR), Ottawa, Ontario, Canada. Haïfa Mahjoub is the recipient of a PhD student scholarship from the International Chair on Cardiometabolic Risk. Philippe Pibarot holds the Canada Research Chair in Valvular Heart Diseases, CIHR. Patrick Mathieu is a research scholar from the Fonds de Recherche en Santé du Québec, Montreal, Québec, Canada. Jean-Pierre Després is the Scientific Director of the International Chair on Cardiometabolic Risk which is based at Université Laval.
Competing interests None.
Patient consent Obtained.
Ethics approval Institutional Review Board of the Quebec Heart and Lung Institute.
Provenance and peer review Not commissioned; externally peer reviewed.
Data sharing statement Additional unpublished data from the study are only available to the corresponding author and co-authors.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.