Background: Abnormalities in body water distribution are common in chronic kidney disease (CKD) patients. Volume
expansion, even in the absence of overt edema, contributes to high blood pressure, and progressive volume expansion
eventually leads to clinical edema and fluid overload. Total body water (TBW) can be accurately estimated by multifrequency
bioelectrical impedance analysis (MF-BIA) which has been proposed for earlier detection of subclinical edema in CKD
patients.
Objective: To study body fluid distribution and edematous states in CKD patients measured by MF-BIA, compared with
clinical edema assessed by physical examination. In addition, to evaluate the correlation of MF-BIA estimated TBW and
anthropometry-derived TBW calculated by Watson formula.
Material and Method: CKD patients at Rajavithi Hospital together with healthy adults were prospectively enrolled during a
12-month period. The body fluid compositions assessed by bioelectrical impedance analyzer (InBody® S20, Republic of
Korea) were taken immediately after physical examination for edema detection. The patients were categorized into stages 1 to
5 according to CKD staging in the NKF-K/DOQI guidelines, and reclassified into 3 groups of stages1-2, stages 3-4, and stage
5.
Results: Sixty-nine CKD patients were compared with 48 healthy volunteers. The estimated glomerular filtration rate (GFR) in CKD patients and normal controls were 53.5+41.1 and 113.9+0.8 ml/min/1.73 m2 respectively. The extracellular water
(ECW) to TBW ratio, which represents edematous state if higher than 0.4, was significantly higher in patients with CKD stages 3-4 (0.400+0.008) and stage 5 (0.404+0.011), than in those in CKD stages 1-2 (0.393+0.009) and controls (0.385+0.007)
(p<0.001). The prevalence of edematous state detected by BIA (edema-BIA) in CKD patients was significantly greater than in
normal controls (78.3% vs. 25.0%, p<0.001). The number of CKD patients with edema-BIA was also significantly higher than the number of patients with clinical edema (36.2%), which represented a significant proportion of patients (42.1%) with
subclinical edema. The sensitivity and specificity of edema detected by physical examination in all CKD patients compared to the assessment by MF-BIA were 44.4% and 93.3% respectively. There was a significant correlation between the TBW calculated by the Watson formula and TBW estimated by MF-BIA (r2 = 0.848, p<0.001).
Conclusion: The present study demonstrated that assessment of body fluid distribution by MF-BIA was a reliable measure.
Subclinical edema actually occurred in early stages of CKD before detection of overt edema by physical examination. TBW
calculated by Watson formula can alternatively be used for evaluation of hydration status and can assist physicians in
prescribing appropriate management for CKD patients.

Keywords: Body fluid distribution, Subclinical edema, Chronic kidney disease, Multifrequency bioimpedance analysis