In a nondiabetic sample of adults aged 70–79 years, CKD was independently associated with a 40% increased odds of IR and was associated with higher visceral fat mass. However, the correlates of IR, including multiple measures of fat mass, were largely similar in those with and without CKD. Only the association of adiponectin with IR varied according to CKD status; specifically, adiponectin was associated with a reduced odds of IR among those without CKD but not among those with CKD. These findings suggest that the mechanisms of IR in CKD are similar to those without CKD.
It has long been known that CKD is associated with IR even in the absence of diabetes [21
]. Several large prospective studies have demonstrated that IR is associated with an increased risk for developing CKD [22
]. CKD may also contribute to IR. The evidence that kidney disease directly causes IR comes mainly from animal models. In experimental models, kidney failure led to hyperglycemia and hyperinsulinemia with glucose loads and fasting hyperinsulinemia [3
]. We were not able to find any prospective studies that evaluated whether kidney disease predicted the development or worsening of IR. IR is present in mild kidney dysfunction and progresses with decreasing GFR [6
] and improves after beginning hemodialysis [28
In the general population, IR is associated with central adiposity [29
]. The accumulation of abdominal fat can occur in either the subcutaneous or visceral compartments and there is controversy as to which type of fat accumulation is more closely associated with IR. Some studies have shown that IR is more closely associated with visceral fat [30
], while others have shown the association to be stronger with subcutaneous fat [32
]. A recent longitudinal study showed that visceral but not subcutaneous adiposity in Japanese Americans predicts future IR [34
]. The mechanism of IR in obesity is thought to involve the disruption of insulin signaling via increases in free fatty acids, adipokines and inflammatory markers, together with reduced adiponectin secretion [35
The mechanism of IR in nondiabetic CKD is not fully understood. Many different mechanisms have been proposed including uremic carbamylation of signaling proteins, acidosis, changes in the apolipoprotein profile, oxidative stress, accumulation of free fatty acids and inflammation [27
]. CKD is associated with a loss of lean mass and a small increase in fat mass [41
] thus BMI and abdominal circumference may be poor surrogate markers for fat mass. In addition, accumulation of adipokines, such as leptin, resistin and visfatin, as GFR declines may also contribute to IR [40
]. Although resistin has been shown to mediate obesity-induced IR in mice, studies do not support a similar role for resistin in humans [10
]. Adiponectin is an adipokine that is made exclusively in adipocytes; it is anti-inflammatory and insulin sensitizing [42
]. In the general population, adiponectin is inversely associated with IR and fat mass. Despite higher levels of IR in end-stage renal disease, adiponectin levels are higher than in healthy controls, which may represent a counter-regulatory response to inflammation or reduced clearance [44
]. Within persons with kidney disease, some studies [11
], but not all, have found adiponectin inversely associated with fat mass or visceral fat [14
]. Thus, there may be more than one opposing determinant for adiponectin levels in CKD (e.g. inflammation leading to higher levels and visceral fat leading to lower levels). Whether this explains the significant interaction seen in this study is not clear.
A strength of the study is that the data include CT-based fat measurements allowing separation of abdominal fat into subcutaneous and visceral fat and measurement of muscle fat and measurement of several cytokines associated with IR. Weaknesses of the study include the lack of actual measurement of GFR as well as the limitations of using HOMA to assess IR. Although the euglycemic clamping method is the gold standard for measuring IR, this technique is not feasible in a large epidemiologic study. Shoji et al.
] did find a strong relationship of HOMA with insulin clamp studies in kidney failure (r
= −0.668). Few Health ABC participants had advanced CKD, thus we were unable to determine whether these associations differed by severity of CKD. It is possible that CKD may contribute to IR only when eGFR is severely reduced. We also did not have information on microalbuminuria, which is commonly associated with IR. The age range is restricted to those aged 70–79 years, who were well functioning, which limits generalizability. Finally, as this was a cross-sectional study, we could not determine whether the observed associations were causal.
This study demonstrates that the correlates of IR, including CT measures of greater fat mass, are similar among elderly individuals with and without CKD, suggesting that the mechanisms of IR may be similar in the CKD population. The similar associations with IR in CKD imply that interventions directed at obesity (e.g. exercise) would have similar benefit as in the general population.