Abstract
Background and Aims
Chronic kidney disease (CKD) patients frequently have an altered body composition driven by metabolic disorders from the uremic syndrome that usually leads to increased protein catabolism, with obesity and muscle impairment being common conditions associated with worse clinical prognosis and high mortality rates. Therefore, with increased mortality and disability rates of CKD patients in the last quarter of a century and the association of a poor body composition with low survival, routine and longitudinal assessment of body composition could improve clinical outcomes. Due to limited availability of reference methods to assess nutritional status, alternative methods are used. In view of the above, our goal was to evaluate the agreement between multifrequency bioelectrical impedance spectroscopy (BIS) and Dual-energy X-ray Absorptiometry (DXA) for assessment of body composition in CKD.
Method
Cross-sectional and prospective analyses by DXA (Hologic, GE®) and BIS (BCM, Fresenius Medical Care®) in whole-body (BISWB) and segmental (BISSEG) protocols were performed in CKD non-dialysis-dependent, hemodialysis and peritoneal dialysis (for at least 3 months), and renal transplantation (for at least 6 months) adult (18 ≤ age ≤ 60 years old) patients. Measurements were performed consecutively by the same professional after an 8-hour fast, drainage of the peritoneal dialysate and just after the midweek hemodialysis session. Intraclass correlation coefficient (ICC) and Bland-Altman plots were evaluated for agreement analysis in group and individual levels, respectively; linear regression analysis was performed for bias assessment and development of new equations; ROC curve was constructed for diagnosis of inadequate error tolerance (DXA - BIS > ± 2kg).
Results
A total of 266 patients were included: 137 men (M) and 129 women (W); 81 were in non-dialysis-dependent treatment, 83 in hemodialysis, 24 in peritoneal dialysis, and 80 had renal transplantation. Total sample had a mean age of 47 ± 10 years old. CKD was secondary to systemic arterial hypertension in 29% of the total sample, to glomerulonephritis in 25%, to diabetes mellitus in 10%, to polycystic kidney in 7%, to glomerulosclerosis and systemic syndromes in 8%, and to other causes and unknown etiology in 20%. Fourteen patients (4 M and 10 W) were in automated and 9 (4 M and 5 W) in continuous ambulatory PD. KTx was by living donor in 18 (14 M and 4 W) and by deceased donor in 63 (34 M and 29 W) patients. The agreement with DXA was greater for BISWB than BISSEG; for fat mass (FM) (ICC in M = 0.89; ICC in W = 0.93) than for fat free mass (FFM) (ICC in M = 0.57; ICC in W = 0.52). Bland-Altman plots showed high limits of agreement (FFM: from -9.51 to 15.64kg; FM: from -7.71 to 7.32kg) with greater bias for FFM as muscular mass increases and for FM in extremes of body fat. The agreement was lower when using the prospective data (body change analysis) (ICC for FFM in M = 0.20; ICC for FFM in W = 0.49; ICC for FM in M = 0.46; ICC for FM in W = 0.58). The factors that interfered in bias between methods were extra to intracellular water ratio (ECW/ICW), body mass index, fat mass index, waist circumference, resistance and reactance (adjusted r2 for FFM = 0.90; r2 for FM = 0.87). FFM had poorer agreement in the last tertile of ECW/ICW sample (ICC in M = 0.69, 0.68 and 0.51; ICC in W = 0.71, 0.74 and 0.38 for first, second and third tertiles, respectively). An ECW/ICW cut-off point of > 0.725 for inadequate error tolerance was determined. New prediction equations for FFM (r2 = 0.91) and FM (r2 = 0.89) presented adequate error tolerance in 55% and 63% in the validation sample compared to 30% and 39% of the original equation, respectively.
Conclusion
For body composition evaluation in CKD, BIS applied using the whole-body protocol, in normal hydration CKD patients is as reliable as DXA; BIS must be used with caution among overhydrated patients with ECW/ICW > 0.725. The newly developed equations are indicated for greater precision.
Evaluation of Body Composition in Hemodialysis Thai Patients: Comparison between Two Models of Bioelectrical Impedance Analyzer and Dual-Energy X-Ray Absorptiometry