Body composition in diabetic subjects with chronic kidney disease: interest of bio-impedance analysis, and anthropometry

Abstract

Background/aims: Lean body mass (LBM) is reduced in uremia, but this has not been reported in diabetic nephropathy.

Subjects and methods: We compared predicted % LBM to DEXA measurements in 10 non-diabetic uremic, 10 non-uremic diabetic and 10 uremic diabetic subjects matched for age, gender and BMI. We also measured % LBM by anthropometry, bio-impedance analysis (BIA) and compared them with DEXA in 49 diabetic subjects with a wide range of renal failure. The results were compared and a Bland & Altman procedure was performed. Associations between glomerular filtration rate (GFR) and % LBM were tested.

Results: In matched groups, predicted % LBM values were overestimated in non-diabetic uremic subjects, and underestimated in non-uremic diabetic subjects. In uremic diabetic subjects, the error was intermediary. As compared to DEXA (% LBM: 69.0 +/- 7.1%), measurement of % LBM by anthropometry (71.4 +/- 8.0%, p < 0.05) and BIA (67.2 +/- 7.6%, p < 0.05) were biased in the 49 diabetic subjects. The mean of anthropometric and BIA (Ant+BIA) were similar to DEXA results (69.3 +/- 6.8%, p = 0.64), with best correlation coefficients and Bland & Altman plots. GFR was correlated to % LBM assessed by DEXA, BIA and Ant+BIA.

Conclusion: In diabetic subjects with chronic kidney disease, LBM should be measured, rather than predicted. A good evaluation is possible, even without DEXA.