Introduction

Kidney stones are a common condition, with an estimated prevalence of 11 % among males and 7 % among females [1]. Recent reports have shown that kidney stones might entail an increased risk of cardiovascular disease; [2] on the other hand, it has been suggested that patients affected with calcium kidney stones tend to have reduced bone mass compared with those without kidney stones. Interestingly, in osteoporosis a similar association with increased risk of cardiovascular events has been observed together with abnormal arterial stiffness [3, 4]. It is tempting to try and streamline the underlying pathophysiologic mechanisms into a single pathway binding together mineral bone leak, kidney stone formation and vascular calcifications [5]. In order to do so, we compared both arterial stiffness, an independent predictor of cardiovascular events, and bone density among subjects with and without idiopathic kidney stones, with the aim of analyzing simultaneously the independent relationship between calcium kidney stones, arterial stiffness and bone metabolism.

Materials and methods

Study participants

We prospectively enrolled 42 patients with recurrent calcium nephrolithiasis that attended the Outpatient Stone Clinic at the Division of Nephrology, University of Verona Hospital. Patients were classified as idiopathic calcium stone formers after exclusion of systemic conditions such as primary hyperparathyroidism, sarcoidosis, intestinal malabsorption, and renal tubular acidosis and medullary sponge kidney. All included patients had stones made of calcium oxalate (defined as >50 % calcium oxalate at chemical analysis of the stone). All were naive of any drug treatments potentially capable to modify blood pressure, the metabolic renal stone risk and mineral bone density.

After written informed consent was obtained, patients undertook a comprehensive medical examination, 24-h urinary collections for metabolic evaluation, fasting blood samples, pulse-wave velocity (PWV) measurements and dual-energy X-ray absorptiometry (DEXA) scan.

The following laboratory tests were performed. In serum: creatinine, calcium, phosphate, 25, OH vitamin D3, alkaline phosphatase, and intact parathyroid hormone (PTH) (HPLC Eureka kit, Eureka, Ancona, Italy).

The patients underwent two 24-h urine collections obtained within a 4–6 weeks interval, following their usual diet. Mean values were calculated for the following variables: volume, pH, calcium, phosphate, citrate (by citrate lyase), and oxalate (by oxalate decarboxylase) [6]. Collections for calcium and phosphate were added 10 mL of 15 % chloridric acid.

Creatinine clearance was calculated from urine and blood concentrations of creatinine and urine volume.

PWV measurements were obtained with PulsePen (Diatecne, Milan, Italy), a non-invasive portable device [7]. The PWV was calculated as distance between the measurement sites divided by a transit time delay between radial and carotid pulse wave and expressed in meters per second (m/s). Abnormal arterial stiffness was arbitrarily defined as any measurement of carotid-radial pulse-wave velocity (CR-PWV), carotid-femoral pulse-wave velocity (CF-PWV) or augmentation index (AI) above the 90th percentile of the sample distribution.

Pulse pressure (PP) was calculated as the difference between systolic and diastolic blood pressure. Mean arterial pressure (MAP) was calculated as the sum of diastolic blood pressure and 1/3 of the PP.

Bone mineral density (BMD; g/cm2) was measured on a level with the vertebrae (L1–L4), femoral neck and trochanter using dual-energy X-ray absorptiometry (DEXA) (Hologic QDR 4500 fan beam densitometer with software version 8.21, Waltham, MA, USA). Bone density data are presented here as site-specific T-scores. We also stratified the study sample according to the presence of reduced bone density, defined as the presence of osteopenia (any T-score <−1 but ≥−2.5) or osteoporosis (any T-score <−2.5).

We prospectively enrolled both stone formers and non-stone formers in the study; however, the final sample was unbalanced with regard to distributions of age and sex, hence, we decided to perform individual matching in order to obtain homogeneous groups, at the cost of excluding a fraction of participants from the analysis. A sample of 42 age- and sex- matched individuals without a history of kidney stones formed the unexposed group. These subjects were enrolled among healthy volunteers from the nursing and medical staff and underwent PWV and DEXA measurements with the same procedures used for the exposed group. Laboratory investigations were not carried out among these unexposed participants.

To avoid residual confounding by variables strongly related with both arterial compliance and bone density, we decided to exclude post-menopausal women as well as smokers and diabetics from the study.

The study adhered to the Declaration of Helsinki and the Faculty Institutional Review Boards approved the protocol of the study.

Statistical analysis

Continuous variables were reported as means and standard deviation (SD) or medians and interquartile ranges (IQR) and categorical variables were reported as frequencies and percentages; independent samples t tests, Wilcoxon tests and Fisher exact tests were used for differences between groups, respectively. CR-PWV, CF-PWV and AI were not normally distributed and were log-transformed.

Hypercalciuria was defined as urinary excretion of calcium >300 mg/24 h for males and >250 mg/24 h for females; hyperoxaluria was defined as urinary excretion of oxalate >45 mg/24 h; hypocitraturia was defined as urinary excretion of citrate <320 mg/24 h.

Abnormal arterial stiffness (AAS) was arbitrarily defined as any measure of arterial stiffness above the 90th percentile of the pooled distribution. Reduced bone density (RBD) was defined as any T-score <−1 SD.

Measures of arterial stiffness (CR-PWV, CF-PWV, AI and AAS) and of bone density (T-scores and RBD) were compared between stone formers and non-stone formers with unadjusted and multivariate adjusted linear and binomial regression models. We reported the following estimates of association from regression models: betas from linear models, geometric mean ratios (GMR) from log-linear models, prevalence ratios (PR) from binomial models; all the estimates were presented with their corresponding 95 % confidence interval (CI). Multivariate adjustment included age, sex and body mass index (BMI); models with measures of arterial stiffness as dependent variables were further adjusted for MAP, PP and heart rate (HR). With the use of mediation regression techniques, we estimated the proportion of the “effect” of kidney stones on arterial stiffness mediated by bone abnormalities and vice versa [8].

The association between AAS, RBD and clinical and biochemical parameters was investigated with regression models adjusted for age, sex and BMI. For these analyses, the p value for significance was Bonferroni-corrected to reflect the number of comparisons and avoid inflation of the type I error.

A p value ≤0.05 was considered statistically significant.

All statistical analyses were performed with R 3.1.0 (R core team, Vienna, Austria) and Stata 12.1 (StataCorp, TX, USA).

Results

Study population

Our sample consisted of 42 idiopathic recurrent calcium stone formers and 42 age- and sex-matched healthy non-stone formers. The baseline characteristics of the study sample according to stone status are shown in Table 1. Matching resulted in the two groups being comparable not only in terms of age and sex distribution, but also of average BMI and blood pressure.

Table 1 Characteristics of the study sample by kidney stones
Full size table

All the participants had measures of arterial stiffness available for analysis; AAS was present in 22 of the 84 participants (24 %).

Sixty-four participants had at least one T-score measurement available for analysis; RBD was present in 35 of the 64 participants (55 %).

Association between kidney stones and arterial stiffness

All the measures of arterial stiffness considered were higher among stone formers compared with non-stone formers (Table 2). In multivariate adjusted models, log-transformed values of PWV-CR, PWV-CF and AI remained higher among stone formers, even if the association was only marginally significant for AI (Table 5). The prevalence of AAS was significantly higher among stone formers compared with non-stone formers (Table 2). The multivariate adjusted regression model confirmed the significance of the association (Table 3).

Table 2 Measures of arterial stiffness in stone formers and non-stone formers
Full size table
Table 3 Association between kidney stones and measures of arterial stiffness
Full size table

Urinary excretion of calcium was higher among those with AAS, with a median (IQR) of 305 (118) mg/24 h, compared with 252 (130) mg/24 h among those without AAS. However, the association was not statistically significant (p = 0.37).

Association between kidney stones and bone density

All T-scores measured were significantly lower among stone formers compared with non-stone formers (Table 4). In multivariate adjusted models, all T-scores remained significantly lower among stone formers (Table 5).

Table 4 Measures of bone density in stone formers and non-stone formers
Full size table
Table 5 Association between kidney stones and measures of bone density
Full size table

The prevalence of RBD was significantly higher among stone formers compared with non-stone formers (Table 4). The multivariate adjusted logistic regression model confirmed the significance of the association (Table 5).

Association between kidney stones, reduced bone density and abnormal arterial stiffness

The prevalence of AAS was higher among those with RBD compared with those with normal bone density (34 vs 14 %, p = 0.08). However, the association became non-significant after adjustment for age, sex and BMI (p = 0.23). Similarly, the mediation regression analysis showed that the amount of “effect” of stone status on AAS mediated by RBD was not significantly different from zero (average percent mediation −0.06, 95 % CI −1.54, 1.26, p = 0.87).

Predictors of abnormal arterial stiffness and reduced bone density among stone formers

All the variables presented in Table 1 were included in age-, sex- and BMI-adjusted analyses for the association with AAS and RBD among stone formers. None of the predictors included in the analysis was significantly associated with either AAS or RBD at the Bonferroni-corrected p value threshold.

Discussion

A number of cross-sectional studies disclosed the possible association of kidney stones with cardiovascular disease [9–12]. Recently, three prospective studies have confirmed that incident stone formers have a 20–60 % higher risk of having incident CV episodes [2, 13, 14]. However, the epidemiological design of those studies had precluded the possibility to ascertain whether the risk is associated with a specific phenotype of renal stone disease.

To obtain insights on this issue and on the liaison with the metabolic bone disease, we compared measures of arterial stiffness and bone density among 42 subjects with recurrent idiopathic calcium oxalate stones and 42 age- and sex-matched non-stone formers. We found that stone formers had significantly higher values for all the measures of arterial stiffness considered: CR-PWV, CF-PWV and AI. PWV is considered as the “gold standard” in the assessment of arterial stiffness, and CR and CF values are thought to reflect different components of the arterial wall, with the former being a measure of the stiffness over central arteries and the latter over peripheral muscular arteries. AI is considered a measure of arterial stiffness as well as PWV, and also provide different information on the status of the arterial wall compared with PWV [15]. All of the above mentioned measures have been suggested as independent predictors of cardiovascular events [16–18]. The consistency of the finding for all the measures analyzed suggests that such process might be systemic and involve central elastic arteries as well as peripheral muscular arteries. The prevalence of AAS, defined in our study as any measure of arterial stiffness above the 90th percentile of the study sample, was higher among stone formers compared with non-stone formers (36 vs 12 %, p = 0.01), and the difference remained significant even after adjustment for potential confounders such as age, sex, BMI, MAP and HR (controlled for by statistical adjustment) and smoke or diabetes (controlled for by restriction), suggesting an independent association between kidney stones and reduced arterial compliance. Even though a relationship with markers of pre-clinical atherosclerosis has been already reported in stone formers [19], to the best of our knowledge this is the first study to compare measures of arterial stiffness among subjects with and without idiopathic calcium kidney stones. Actually, in the CARDIA study the stone condition was ascertained only by the referral of a stone episode, without any knowledge on the nature of the stone disease. In a parallel study, we investigated measures of arterial stiffness among patients with secondary forms of calcium nephrolithiasis (medullary sponge kidney and distal renal tubular acidosis) and did not find any significant association with abnormal arterial stiffness (unpublished data), suggesting that this condition might be specific of idiopathic calcium nephrolithiasis and that the reported epidemiological association with cardiovascular outcomes might in fact be driven by this frequent population. Previous studies that reported an association with cardiovascular outcomes or markers thereof did not analyze whether the association was driven by specific nephrolithiasis phenotypes.

We found a significantly reduced bone density among stone formers compared with healthy participants. Stone formers had significantly lower values for all the T-scores and a higher prevalence of RBD, defined as the presence of either osteopenia or osteoporosis (83 vs 21 %, p < 0.001). The relationship between kidney stones and RBD remained significant even after adjustment for age, sex and BMI. Our study confirms previous findings of abnormal bone loss among stone formers: a relationship between kidney stones and the prevalence or risk of vertebral fractures has been long known [20, 21]. It has also been reported that lithiasic activity is associated with markers of bone remodeling [22]. Participants with RBD did show an almost significantly higher prevalence of AAS (34 vs 14 %, p = 0.08), however the association disappeared after adjustment for confounders. Moreover, adjustment for RBD did not significantly change the association between kidney stones and abnormal arterial stiffness, and in fact the mediation analysis showed that a negligible amount of the “effect” of kidney stones on arterial stiffness would be mediated by bone density, suggesting that bone loss is not in the causal pathway between kidney stones and alterations of the vascular wall, although the lack of association might be due to the relatively small sample size of our study. We also could not show any role of disease characteristics such as disease duration or number of stone episodes on RBD. The lack of an association between either blood or urine levels of calcium and AAS or RBD would further confirm that two separate processes might link kidney stones on one hand and arterial stiffness and osteopenia on the other hand, standing against the hypothesis that calcium leak from the tubule would mobilize calcium from the bone and into the vascular wall. However, since our study has a cross-sectional design, we cannot exclude that there might in fact be a single pathophysiologic process that might develop over time, hence making it difficult to observe both the vascular and the bone alterations in the same patient simultaneously. Future adequately powered longitudinal prospective studies are warranted to better understand this relationship. Another limitation of our study is that the unexposed group was composed of health professionals, hence potentially reducing the generalizability of our findings; however, characteristics such as BMI and blood pressure were similar in the two groups; furthermore, the unexposed participants were matched by age and sex, making differences in distributions of unmeasured confounders less likely.

In conclusion, the study shows that idiopathic calcium stone formers have an abnormal arterial stiffness, a surrogate marker of atherosclerosis. This is in accordance with epidemiologic studies indicating that idiopathic calcium renal stone formers, the great majority of renal stone patients, are at increased risk of cardiovascular events. The data also suggest that results obtained by previous cohort studies disclosing an association with increased carotid intima-media thickness, and CV endpoints, respectively [2, 13, 14, 19], are probably explained by the higher CV risk imparted by idiopathic calcium stones rather than by other special kind of stone formers, for example the urate stone patients.