Cardiovascular disease is the most important cause of mortality in chronic kidney disease (CKD). Keith et al.1 reported that this population presented a greater risk of dying as a result of a cardiovascular event than of starting dialysis. In recent years, different research has focused on the search for new instruments that identify this risk as opposed to the traditional risk scales that are not valid in CKD. The carotid ultrasound and the ankle-brachial index (ABI) have proven their usefulness as subclinical atherosclerosis screening markers in these patients.2 Nevertheless, both explorations require specific technological equipment that is not always available in routine clinical practice.

On the other hand, access to the simultaneous measurement of triglyceride levels and the waist does not call for a major diagnostic effort and renders it possible to define the hypertriglyceridaemic waist phenotype (HWP). The HWP has proven its usefulness in the general population for the detection of the so-called atherogenic metabolic triad (AMT): hyperinsulinemia, hyperapolipoproteinaemia B and elevation of small dense low-density lipoprotein-cholesterol particles (sdLDL-C).3 It is also a useful screening tool for coronary disease and cardiovascular mortality.4,5

Zhe et al.6 identified a relationship between HWP and carotid intima-media thickness (CIMT) in the population with CKD. However, we know that this disease involves a greater comorbidity than that of the population in which the HWP was initially defined, with a complex pathophysiology for the development of cardiovascular disease and differential characteristics in relation to lipid metabolism and obesity, as well as a tendency to undergo changes in body composition. For this reason, our hypothesis poses the existence of different HWP cut-off points in subjects with CKD. We have not identified any studies that have evaluated this possibility before. This work will seek to identify the HWP cut-off points that best detect an increase in cardiovascular risk (CVR) in the population with CKD.

With regard to the cut-off points selected to define HWP in the Europids population (triglycerides above 177 mg/dl in men or 133 mg/dl in women, together with a waist greater than 90 cm in men or 85 cm in women),3,10 in this work we observed that levels above 143 mg/dl combined with a waist greater than 102 cm in men or 94 cm in women render it possible to optimise this tool’s diagnostic performance in the population with CKD. This variation could have several explanations. First of all, unlike the general population, where the cut-off points were selected by identifying non-traditional CVR factors (those that define AMT), we used the presence of a subclinical, but already established, atherosclerotic lesion as an outcome variable. Moreover, our kidney patient cohort was older and presented greater comorbidity compared to that of the original studies (median age 43.1 years, non-obese, with no treatment for coronary disease, diabetes, dyslipidaemia or endocrine disorders).3,10

The prevalence of HWP was considerable, irrespective of the definition used. The specific definition shows an age-related increase, as has also been described previously in the population without CKD,11,12 with the greatest peak in post-menopausal women. Other authors have previously shown an increase in visceral fat in women with this condition, relating it to a reduction in serum levels of oestrogens and dehydroepiandrosterone sulphate.13,14

We identified an association between the presence of this phenotype and different CVR factors. Thus, as in studies in a population without CKD,3,10 we observed a significant association with the atherogenic metabolic triad, including the increase in apolipoprotein B levels and insulin resistance. Although we did not have a determination of sdLDL-C levels, we observed that they were associated with higher levels of non-HDL cholesterol, a parameter that presented a good correlation with it (r = 0.76).15 We also identified an association with elevated levels of us-CRP to limits deemed to constitute a greater CVR (>2 mg/L).16 Individuals with HWP presented higher systolic blood and pulse pressure levels, suggesting greater arterial stiffness, although we did not detect any differences in vascular calcification as assessed by the presence of an ABI > 1.2. Similar to what has been described in obese subjects, the HWP was associated with a reduction in levels of 25OHD.3. Taking into account the poorer prognostic outcomes associated with levels <30 ng/dL in patients with CKD,17 we wondered whether screening and supplementing could be particularly beneficial in these patients.

The regression analysis showed an independent association between specific HWP and the presence of AD2-3. In relation to the variables selected to adjust the model, it should be emphasised that some of them were selected from the literature review, including previous publications with the NEFRONA cohort.9,18–21 The aforementioned work involved a noteworthy assessment of risk factors associated with the presence of atheroma plaque, pathological ABI or the progression of AD in patients with CKD. On this basis, we selected variables such as levels of phosphorus associated with greater CVR according to gender (above 3.5 mg/dl in men and 5 mg/dl in women)20 or the significant effect of the interaction between the levels of ferritin and age.21

Similarly, the specific HWP behaved as an independent risk factor for the development of CVEoM. However, we observed that this effect depended on the interaction with phosphorus serum levels and was only significant in patients with phosphorus below a 3.5 mg/dl at the beginning of the study. In the stratified analysis by CKD stages, the interaction only remained significant in the G4-5 stages. Garagarza et al.22 showed an association between hypophosphoremia and increased mortality in advanced forms of CKD, stating that the presence of hypoalbuminemia and hypovolaemia acted as possible confounding variables in these patients. In our work, we were unable to evaluate this information at the end of follow-up, although initial albumin values did not modify the outcomes obtained. Moreover, the possible role of the adipokines in these associations must be considered. Thus, in the population with CKD, the reduction in GF is described as the main determinant of elevated levels of adiponectin, which increase logarithmically as of stage 4, whereas body mass index is the main determinant of elevated leptin.23 By both routes, patients in G4-5 stages with HWP could present an elevation of both adipokines. More recently, the influence of adipose tissue in bone and mineral metabolism regulation has been described. In obese patients, the adipose tissue increases the release of factors associated with the stimulation of FGF-23, such as interleukin-6, tumour necrosis factor alpha and leptin. Moreover, experimental studies in transgenic mouse models with over-expression of adiponectin show how a surplus of phosphorus in the diet is associated with elevated FGF-23 levels and phosphorus excretion. It is no coincidence that FGF-23 is related to the development of structural cardiopathy and cardiovascular disease in the population with CKD.24–26 The association between HWP and CVEoM identified in subjects with lower phosphorus could be related to these physiopathogenic mechanisms. However, in our analysis we lacked sufficient information to be able to confirm this hypothesis.

The multivariable regression analysis showed a significant phenotype effect which also extended to the population on haemodialysis. In this same line, Postorino et al.27 studied patients on haemodialysis, describing a progressive increase in cardiovascular mortality and for all causes from fixed surpluses of triglycerides of 50 mg/dl in patients with a waist greater than 95 cm, as well as a reduction of the same as triglycerides increase in patients with lower waist limits. Therefore, our results, together with these findings, point to the need to consider a specific definition in the population on haemodialysis.

The main strong points of the study are those of the NEFRONA cohort, with a population sample that was representative of the entire Spanish territory, as well as information on a broad number of variables that enabled us to perform a powerful search and subsequently adjust for numerous confounding factors. Similarly, the NEFRONA Project featured a rigorous methodological design.7–9 Finally, attention should be drawn to the absence of cardiovascular events prior to the recruitment of the patients in the study, allowing us to identify cut-off points that could be useful in the detection of increased CVR in the subclinical stages.

This work also presents certain limitations. The first one is that we used data obtained transversely to infer a longitudinal relationship. Besides the models that we selected, other possible confounding variables could be involved, so other prospective studies that establish a more precise casual relationship would be called for. Secondly, the study needs to be completed with an external validation analysis. Thirdly, the group of subjects in the G1-2 and A > 1 stage was small, although we sought to control this bias by applying the Bonferroni correction to the analyses. Finally, in view of the absence of kidney transplant patients in the cohort and the added exclusion of patients on peritoneal dialysis due to the absence of a waist record, our results cannot be main-streamed to either of these two subpopulations. In this regard, there is work currently ongoing that establishes a suitable waist measurement technique in the population on peritoneal dialysis and relates this treatment to a progressive increase in this measurement, in turn independently associated with the reduction in adiponectin levels.28,29 At the same time, several publications relate inflammatory status, risk of diabetes and kidney transplant cardiovascular prognosis to abdominal circumference and other visceral obesity markers.30–32 For all these reasons, it would be very interesting to be able to study HWP in both populations.

In conclusion, adjusting the limits of the definition of HWP would appear to be necessary in the CKD population. The specific HWP of this population constitutes a simple tool that can be useful in consultations to identify a group of kidney patients with an additional CVR, and this also includes the population on haemodialysis. Our results arouse interest in the development of studies to evaluate the role of visceral obesity in the regulation of bone mineral metabolism in patients with CKD, as well as intervention design studies to evaluate possible interventions aimed at improving the cardiovascular prognosis of these patients.

The cut-off points for defining HWP in the population with CKD presenting the highest diagnostic performance were triglyceride levels >143 mg/dl and waist >102 cm in men and >94 cm in women.

The differences observed with regard to the cut-off points described in the literature could be accounted for by the greater age and comorbidity of the population with CKD, as well as the use of a different diagnostic variable (subclinical, but established, atherosclerotic disease, compared to atherosclerotic disease risk factors: atherogenic metabolic triad).

The prevalence of HWP in CKD is considerable, and also in patients on haemodialysis.

The interaction identified between HWP and phosphorus levels with regard to the development of CVEoM arouses interest in extending the study of the relationship between this phenotype and bone and mineral metabolism regulation.

The authors have no conflicts of interest to declare.