Association of serum calcium and heart failure with preserved ejection fraction in patients with type 2 diabetes

Background Type 2 diabetes mellitus (T2DM) is a recognized trigger factor for heart failure with preserved ejection fraction (HFpEF). Recent studies show that higher serum calcium level is associated with greater risk of both T2DM and heart failure. We speculate that increased serum calcium is related to HFpEF prevalence in patients with T2DM. Methods In this cross-sectional echocardiographic study, 807 normocalcemia and normophosphatemia patients with T2DM participated, of whom 106 had HFpEF. Multinomial logistic regression was carried out to determine the variables associated with HFpEF. The associations between serum calcium and metabolic parameters, as well as the rate of HFpEF were examined using bivariate linear correlation and binary logistic regression, respectively. The predictive performance of serum calcium for HFpEF was evaluated using the area under the receiver operating characteristic curve (AUC). Results Patients with HFpEF have significantly higher serum calcium than those without HFpEF. Serum calcium was positively associated with total cholesterol, triglycerides, low-density lipoprotein cholesterol, serum uric acid, HOMA-IR and fasting plasma glucose. Compared with patients in the lowest serum calcium quartile, the odds ratio (OR) for HFpEF in patients in the highest quartile was 2.331 (95 % CI 1.088–4.994, p = 0.029). When calcium was analyzed as a continuous variable, per 1 mg/dL increase, the OR (95 % CI) for HFpEF was [2.712 (1.471–5.002), p = 0.001]. Serum calcium can predict HFpEF [AUC = 0.673, 95 % CI (0.620–0.726), p < 0.001]. Conclusions An increase in serum calcium level is associated with an increased risk of HFpEF in patients with T2DM.


Background
Heart failure (HF) is magnified in individuals with type 2 diabetes mellitus (T2DM), in whom incidence rates are 2-5 times greater than those in the general population [1,2]. Heart failure with preserved ejection fraction (HFpEF) constitutes approximately 50-55 % of the HF population [3] and the prevalence of HFpEF is rising at a rate of around 1 % per year [4], so it is predicted that HFpEF will become the most prevalent phenotype of HF over the next decade [4,5]. Despite robust evidence of prognostic benefit using therapies with angiotensin-converting enzyme inhibitors, angiotensin-1 receptor blockers and β-blockers in heart failure with reduced ejection fraction (HFrEF), all outcome trials in HFpEF to date have failed to demonstrate survival benefit [3,6]. Since much less is known about the pathophysiology and treatment of HFpEF in contrast to HFrEF [5,6], screening potential risk factors in the progression of HFpEF in diabetic patients is of particular importance. is independently associated with increased risk of T2DM [9,10] and cardiovascular disease [11] even in normocalcemic populations.
Based on these findings, we speculate that an alteration in serum calcium is associated with HFpEF prevalence, and we conduct a cross-sectional study to evaluate relationships between serum calcium levels and HFpEF in T2DM patients.

Participants
A total of 807 subjects (463 men and 344 women) were included in this study. We recruited consecutive subjects aged 40 years or older who visited Renmin Hospital for education, evaluation, or treatment of T2DM from 2012 to 2015.
To minimize the possibility that some abnormal conditions may influence the results, patients with any of the following conditions were excluded: (1) history of left ventricular ejection fraction (LVEF) <50 % at any time; (2) isolated right heart failure due to pulmonary disease; (3) dyspnoea due to non-cardiac causes such as pulmonary disease, anaemia, or severe obesity; (4) primary valvular or myocardial diseases, atrial fibrillation, coronary artery or cerebrovascular disease needing revascularisation within 3 months; (5) serum creatinine >130 μmol/L (normal range: 50-130 μmol/L) or urine albumin per gram urine creatinine (Alb/Cr) >300 mg/g; (6) uncontrolled thyroid diseases, history of parathyroid disease or vitamin D-related disorders; (7) medication history including vitamin D, bisphosphonate, estrogen replacement therapy and diuretics which may influence calcium metabolism within the past 1 month; (8) serum calcium out of normal range from central laboratory of Renmin hospital (8.42-10.42 mg/dL, or 2.10-2.60 mmol/L); (9) serum phosphate out of normal range from central laboratory of Renmin hospital (3.00-4.50 mg/dL, or 0.97-1.45 mmol/L).
HFpEF was diagnosed according to the European Society of Cardiology guideline [12]: (1) presence of symptoms and/or signs of HF; (2) LVEF ≥50 %; (3) NT-proBNP >125 pg/mL. T2DM was diagnosed by the American Diabetes Association guideline [13]. Obesity was defined as body mass index (BMI) ≥28 kg/m 2 according to Chinese standard [14]. Dyslipidemia was defined as HDL-C <1.04 mmol/L, LDL-C ≥4.14 mmol/L, or TG ≥2.26 mmol/L [15]. Smoking was defined as "ever smoked" as compared to "never smoked". Hypertension was defined as systolic blood pressure (SBP) ≥140 mmHg and/or diastolic blood pressure (DBP) ≥90 mmHg, or current antihypertensive therapy. Micro-albuminuria was defined as Alb/Cr between 30 and 300 mg/g, and macroalbuminuria was defined as Alb/Cr >300 mg/g. This study was approved by the ethical review board of Renmin Hospital and complied with the Helsinki declaration. Written informed consent was obtained from all participants.

Data analysis
Continuous variables were presented as mean ± standard deviation (SD), as well as frequencies and percentages for categorical variables. Normal distribution was checked by Kolmogorov-Smirnov Test. HOMA-IR and Alb/Cr were logarithmically transformed to approximate normal distribution for analysis. Differences in normally distributed variables were determined by independentsamples T test or One-way ANOVA. If data were nonnormally distributed or not met the homogeneity of variances, a nonparametric test was performed. Chi square tests were applied for categorical variables. Bivariate linear correlation (Pearson correlation) analysis was carried out to evaluate the associations between albumin-adjusted serum calcium and metabolic parameters. Backward stepwise multinomial logistic regression analysis was carried out to determine the variables associated with HFpEF and to estimate confounding factors possibly disturbing the relationship between serum calcium and HFpEF. Binary logistic regression analysis was performed using HFpEF as the dependent variable to analyze the association between serum calcium and HFpEF after adjusting for potential confounders. Odds ratios (OR) with 95 % confidence intervals (CI) were calculated for the relative risk of increased serum calcium level with HFpEF. The ability to predict HFpEF of albumin-adjusted serum calcium was evaluated using the area under the curve (AUC) in the receiver operating characteristic (ROC) curve. All statistical analysis were performed using Statistical Product and Service Solutions (SPSS) version 19.0. All tests were two-sided, p < 0.05 was considered statistically significant.
Significant differences in albumin-adjusted serum calcium (8.96 ± 0.36 vs. 9.22 ± 0.44 mg/dL, p < 0.001) were observed between non-HFpEF and HFpEF group ( Table 1). The patients with HFpEF had higher levels of NT-proBNP, lg HOMA-IR, serum uric acid, serum creatinine and lg Alb/Cr, longer duration of diabetes, greater percentage of female gender, micro-albuminuria and LVH (concentric hypertrophy, especially), as well as lower levels of serum albumin than those without HFpEF.

LV geometry
Compared to subjects in albumin-adjusted serum calcium quartile 1, those in quartile 4 had significant lower percentage of normal LV geometry (31.8 vs. 53.6 %); by contrast, percentage of the subjects with LV eccentric hypertrophy and concentric hypertrophy increased sharply from 11.6 to 20.7 %, 11.1 to 31.8 %, respectively (Table 3).
LAD, IVSD, PWTD, RWT, and LVMI of subjects in quartile 4 were significantly higher than those in quartile 1. As an indicator of systolic function, LVEF levels had no significant differences among groups categorized by albumin-adjusted serum calcium. E/A as well as DT, indicators of diastolic function, had significantly overall downward (1.14 ± 0.20 to 1.03 ± 0.23) and upward (190.39 ± 61.03 to 246.88 ± 49.98) tendencies, respectively, from quartile 1 to quartile 4 ( Table 3).

HFpEF
From albumin-adjusted serum calcium quartile 1 to quartile 4, percentage of the individuals with HFpEF increased sharply from 5.3 to 22.7 % (Table 3).
The binary logistic regression analysis (Table 5) showed the OR (95 % CI) for HFpEF according to changes in albumin-adjusted serum calcium concentration when calcium was a categorical variable (quartiles) or a continuous variable (per 1 mg/dL). In contrast to subjects in quartile 1 (8.42-8. To evaluate the predictive performance of albuminadjusted serum calcium for HFpEF, the AUC in ROC curve was calculated, which was 0.673 [95 % CI (0.620-0.726), p < 0.001] (Fig. 1).

Discussion
Epidemiological studies have associated T2DM with HFpEF [5,21,22]. On the one hand, T2DM is a wellknown trigger factor for HFpEF which exerts important effects on ventricular relaxation/stiffness [5] and coronary microvascular function [23]; on the other hand, diabetic cardiomyopathy is manifested by HFpEF other than HFrEF at an early stage [22]. Given that elevated serum calcium levels are associated with increased risks of T2DM [9,10], an important question arises whether elevated serum calcium contributes to HFpEF prevalence in T2DM.
To the best of our knowledge, this is the first analysis of the relationship between changes in serum calcium levels and the risk of HFpEF that focused specifically on T2DM patients with normocalcemia and normophosphatemia. Our results showed a clear association between the elevated albumin-adjusted serum calcium levels and the increased risk of HFpEF. Such an association is independent of the effects of age, gender, obesity, smoking, hypertension, dyslipidemia, LVMI, HbA1c, Alb/Cr, HOMA-IR, and serum uric acid.
In our study, patients with HFpEF had significantly higher levels of albumin-adjusted serum calcium than those without HFpEF. On the other hand, patients in the highest serum calcium quartile had significantly greater percentage of HFpEF than those in the lowest quartile. Previous studies have demonstrated that higher serum calcium levels are associated with greater risks of incident HF Italic values represent p < 0.05 [24], worse outcomes of HF [25], poorer clinical response to maximization of HF therapy [26]. However, the relationship between serum calcium and HFpEF is unknown. Our study indicates that elevated serum calcium though in normal range is related to HFpEF prevalence in T2DM.
Elevated serum phosphate concentrations have been associated with cardiovascular events including heart failure through its interactions with parathyroid hormone, vitamin D, and fibroblast growth factor 23 in some studies [27,28]. In contrast, the third National Health and Nutrition Examination Survey showed that factors determining serum phosphate concentrations are largely unknown and previously observed associations of serum phosphate concentrations with cardiovascular events are unlikely to reflect differences in traditional cardiovascular risk factors [29]. Furthermore, serum phosphate have been more likey associated with HFrEF and eccentric hypertrophy rather than HFpEF and concentric hypertrophy [30,31]. In line with these studies, our data do not support an association between serum phosphate and HFpEF.
LVH is associated with increased HF risk [22,32]. Though eccentric hypertrophy can occur in HFpEF, ours and previous studies [3,32] demonstrate that concentric hypertrophy is the common form of left ventricular structural abnormality observed in these patients. Furthermore, in patients with T2DM, serum calcium is associated with an increased risk of LVH [33]. In accord with these, patients in our study with concentric hypertrophy had a higher but not significant level of serum calcium (9.13 ± 0.40 vs. 9.05 ± 0.42 mg/dL, p = 0.083) and a greater percentage of HFpEF (28.7 vs. 16.7 %, p = 0.014)  than those with eccentric hypertrophy; however, after adjusted for LVMI (the index for LVH), serum calcium remained significantly associated with HFpEF (Table 5). Hence, the association between serum calcium and HFpEF in patients with T2DM can somewhat, but not fully, be explained by LVH. As a key pathophysiological mechanism of T2DM, insulin resistance is not only accompanied with an increase in intracellular calcium [34,35], but also positively correlated with serum calcium level in ours and other studies [33,36,37]. Meanwhile, there is increasing awareness regarding the associations of insulin resistance with myocardial diastolic dysfunction, cardiomyopathy and heart failure [21,38,39]. In line with these studies, our results showed a significant correlation between the elevated lg HOMA-IR value and the increased risk of HFpEF [OR (95 % CI) = 2.693 (1.366-5.310), p = 0.004].
In the current study, the binary logistic regression analysis showed a significantly association between albumin-adjusted serum calcium and HFpEF. Moreover, consistent with previous studies, our final model using   [3,5], female gender [3,5], uric acid [40], and HOMA-IR [21,38,39] were also related to an increased risk of HFpEF. The potential mechanisms underlying association between serum calcium and HFpEF remains unclear; however, there are some possibilities. On the one hand, the elevation of serum calcium appears to function as a connecting link among various metabolic disorders. Ours and previous studies [33,36,37,[41][42][43] have demonstrated that serum calcium level was positively and linearly associated with glucolipid metabolic parameters including FPG, HOMA-IR, uric acid, TG, TC and LDL-C. Given that various metabolic abnormalities, such as diabetes [3,5], obesity [3,5], hyperuricemia [40], insulin resistance [21,38,39], and metabolic syndrome [21,39,44] have been reported in association with abnormal left ventricular diastolic function, which is the fundamental physiopathologic mechanism responsible for the development of HFpEF [45], the increased serum calcium level may correlate with HFpEF prevalence through metabolic abnormalities. On the other hand, serum calcium level has close relationship with some recognized pathological mechanisms of HFpEF [3] such as LVH [33] and vascular stiffness [11,46]. Patients with HFpEF have a predominant abnormality in left ventricular distolic function [5], which is sensitive to disorders in calcium metabolism [47]. The increased diastolic tension is a result of elevated cytosolic diastolic calcium [8]. Abnormal calcium homeostasis is not only one of the mechanisms in HFpEF [8], but also a prominent feature in the transition from cardiac compensatory hypertrophy to heart failure [48].
Several limitations of this study should be noted. First, in our study no serum parathyroid hormone, vitamin D and fibroblast growth factor 23 levels are available for most of the patients which may help to delineate the underlying mechanisms for the association between serum calcium and HFpEF, so it is impossible to absolutely exclude potential confounding factors including primary hyperparathyroidism and secondary hyperparathyroidism due to vitamin D deficiency and/ or renal insufficiency. To minimize these possibilities, we excluded individuals with serum calcium or phosphate levels outside the reference range. In addition, secondary hyperparathyroidism cannot account for the higher rate of HFpEF among patients with higher serum calcium in our study, because serum calcium levels are well-known lower or low-normal in individuals with secondary hyperparathyroidism. Second, the results are based on single serum calcium measurements; therefore, time course of changes in calcium is not available. Third, the majority of participants in our study were old Chinese, which may limit the generalizability of our results to other age-groups or ethnicities. Fourth, the sample size in this study is only moderate. Finally, the hospital-based cross-sectional study is vulnerable to sample selection bias and cannot establish a cause-effect relationship.

Conclusions
Our results support the reported correlation between calcium and glucolipid metabolism, and extend previous findings of the association between serum calcium and cardiovascular disease, especially heart failure. The increased albumin-adjusted serum calcium level, within the physiological ranges, is independently associated with HFpEF prevalence in patients with T2DM.