Background
Albuminuria is not only an independent prognostic marker for kidney disease, but it is also an established predictor of cardiovascular disease and all-cause mortality [
1‐
3]. However, there has been considerable debate over the threshold level of urinary albumin excretion (UAE) that predicts adverse clinical outcomes. Recently, meta-analyses of general- and high-risk population cohorts have demonstrated that increased UAE, even within the normoalbuminuric range, especially a urinary albumin-creatinine ratio (UACR) of 10–29 mg/g, is an independent predictor of cardiovascular disease and all-cause mortality [
4,
5]. These noteworthy reports on the prognostic significance of the higher range of normoalbuminuria have challenged the previous definition of normoalbuminuria. Furthermore, given the high prevalence of the greater range of normoalbuminuria in the general population [
6], we believe that exploring its risk factors might have important public health and clinical implications.
Vitamin D deficiency is an important public health issue because of its association with numerous adverse conditions [
7]. The association of vitamin D with albuminuria has been suggested; however, most of these studies have been limited to subjects with chronic kidney disease or diabetes, which are susceptible to albuminuria [
8‐
12]. Recently, general population studies have shown that vitamin D deficiency was associated with prevalent microalbuminuria and that vitamin D deficiency posed an increased risk of
de novo microalbuminuria [
13,
14]. However, these studies did not consider high-normal UAE as a separate outcome variable, despite its obvious clinical significance and high prevalence. Moreover, previous studies did not consider a possible non-linear relationship between 25-hydroxyvitamin D [25(OH)D] and UACR levels, although recent large-scale studies have suggested a non-linear relationship between the risks of various clinical conditions and serum 25(OH)D levels [
15‐
20].
Based on previous findings regarding the association between vitamin D and albuminuria, we postulated that low vitamin D levels might be a potential risk factor for high-normal UAE. In the present study, we examined the relationship between serum 25(OH)D and UACR levels in a large number of subjects who underwent voluntary health screenings. We aimed to estimate the threshold level of serum 25(OH)D less than which UAE would begin to increase.
Discussion
In the present study, conducted in a large number of health-screening subjects, we demonstrated a non-linear relationship between serum 25(OH)D and UACR levels in normoalbuminuric subjects, and we identified the threshold level of serum 25(OH)D that predicted an increase in UAE, independent of confounding variables. Because the serum 25(OH)D threshold was lower than the generally accepted lower limit of the normal range, the present study indicates that severe vitamin D deficiency could cause an increase in UAE. Additionally, we demonstrated that the risk of high-normal UAE (UACR 10–29 mg/g) increased as the serum 25(OH)D level decreased; however, the trend was not linear. Thus, the present study raises the important issue that vitamin D could be a potential therapeutic option for the management or prevention of increases in UAE in normoalbuminuric subjects with severe vitamin D deficiency.
Increases in UAE, even within the normoalbuminuric range, constitute a well-known risk factor for adverse cardiovascular outcomes in high-risk individuals [
27,
28]. More recently, high-normal UAE has become a critical public health issue because high-normal UAE has been associated with the risk of cardiovascular disease and all-cause mortality in the general population [
4,
5,
29]. Additionally, because large-scale, population-based studies have documented a high prevalence of high-normal UAE, ranging from 13% to 23% [
30‐
33], its public health aspects require greater attention. However, despite its significance for public health, there have been few studies investigating its causative or potentially modifiable risk factors, other than traditional ones, such as diabetes, hypertension, or obesity. Given that low-grade albuminuria reflects generalized vascular dysfunction via impaired endothelial function, activation of the RAS, and chronic low-grade inflammation [
34‐
37]. conditions relevant to the development of these pathogenic processes could constitute potential risk factors for high-normal UAE. In this respect, vitamin D deficiency is a likely mechanism for the increase in UAE, as it has been shown to play a crucial role in vascular dysfunction. Recent studies have suggested that flow-mediated dilatation, which is a standard tool for measuring endothelial function, was significantly impaired in subjects with vitamin D deficiency [
38,
39]. Vitamin D deficiency also adversely affected vascular function by inappropriate activation of the RAS [
40], and chronic inflammation [
41]. These processes might, in turn, increase UAE, even in the absence of evident kidney disease. The data presented show that severe vitamin D deficiency was significantly, but non-linearly, correlated with high-normal UAE, even after adjustment for traditional risk factors for albuminuria. Similar to our results, a previous study found that low serum 25(OH)D levels were associated with an increased prevalence of albuminuria in a general population [
14]. However, in that study, microalbuminuria (defined as UACR of 17–249 mg/g for men and 25–354 mg/g for women), but not high-normal UAE, was considered to be a dependent variable, so the reference group included a large number of subjects with high-normal UAE, which is a proven risk factor for adverse outcomes.
Recently, mounting evidence has suggested that the inverse associations between serum 25(OH)D levels and risk factors in cardiovascular outcomes might be non-linear, involving a threshold effect [
15,
18,
19]. Given the close pathophysiological association between albuminuria and cardiovascular disease, serum 25(OH)D levels could also have a threshold effect on levels of UAE. The present results of GAM, which is a non-parametric regression method used to address non-linearity between variables [
26], are consistent with the hypothesis that the association between serum 25(OH)D and UACR levels is non-linear. However, the threshold serum 25(OH)D concentration in this study (8 ng/mL) was much lower than that reported in previous studies that explored the associations between serum 25(OH)D levels and risk factors for cardiovascular disease. Previous studies have demonstrated that the risk for cardiovascular disease began to increase significantly at less than 15 to 30 ng/mL of serum 25(OH)D [
15,
18,
19], which is relevant to the current definition of vitamin D deficiency. Considering the high prevalence of vitamin D deficiency in the general population, a substantial number of subjects would be expected to be at vitamin D deficiency-related risk for cardiovascular disease. However, most randomized, controlled trials with vitamin D supplementation have failed to demonstrate improved vascular function or cardiovascular outcomes [
42‐
44], suggesting that low vitamin D levels simply reflect health status, rather than directly causing disease. An alternate explanation for the discrepancy is that vitamin D plays an independent role in the development of cardiovascular disease or vascular dysfunction, when its levels decrease to less than very low levels. This hypothesis was supported by a recent large-scale, population-based study, which demonstrated that only a very low vitamin D level (<5
th percentile of serum 25(OH)D) significantly increased the risk of fatal ischemic heart disease after adjustment for traditional cardiovascular risk factors [
39]. In accordance with the previous study, the present study also demonstrated that the 25(OH)D threshold levels decreased after adjustment for the traditional cardiovascular risk factors in GAM analysis. Additionally, only the lowest category of serum 25(OH)D was correlated with elevated odds of high-normal UAE after multivariate adjustment. These findings suggest that vitamin D levels related to the pathogenesis of low-grade albuminuria might be lower than the lower limit of the normal range of vitamin D. This outcome could be one of the possible indirect explanations for the negative results of randomized, controlled trials regarding the effects of vitamin D supplementation on vascular function, as low-grade albuminuria could reflect generalized vascular dysfunction.
There were several limitations to the present study. First, UAE was measured from only a single random urine sample. Individual variability in UACR levels could have caused measurement error for UAE. Second, there was a lack of information on vitamin D dietary intake and outdoor physical activities, which exert effects on vitamin D status. Third, the current investigation was cross-sectionally designed. We could not establish a causative or temporal relationship between vitamin D levels and UAE. Thus, we could not demonstrate whether use of vitamin D could reduce urinary albumin loss in a healthy population with very low degrees of UAE. Finally, the laboratory test methods were not consistent among the institutions. Thus, to combine laboratory test results from different institutions, a normalization process, which transformed different distributions into identical distributions in statistical properties, was indispensable. We used quantile normalization, which is the most widely used normalization method for genetic studies [
22]. This method makes the distribution of different datasets as similar as possible while conserving the meaning of original values. To consider the reference range of laboratory tests, we partitioned each datasets into three groups (upper range, within range and below range), and each of them were independently normalized. By doing this, we could normalize the datasets while conserving the clinical meaning and considering the reference range. Because this method does not assume any parametric distributions, the method can be applied to any laboratory test results with various distributions.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made.
The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
YJO, RWP, and DKK conceived of the design of this research, and helped to draft the manuscript. DKK supervised this project. YJO, HRJ, HK, HL, NJH, and SKP collected the data. YJO, DY, MK, and SSH analyzed the data. RWP and KWJ interpreted the results. YSK and CSL gave conceptual advice and comments on the manuscript. All authors read and approved the final manuscript.