Body mass index in patients with amputations on peritoneal dialysis: error of uncorrected estimates and proposed correction.

UNLABELLED: "Weight-height" indices including percent of ideal weight (%IW) and body mass index (BMI) are used to estimate degree of obesity in populations and are predictors of survival in dialysis patients. Amputation affects the relationship between weight and height independently of the degree of obesity. Corrections of both %IW and BMI for amputation have been published, but a National (U.S.) computer nutrition program used in the authors' institution uses only the correction for %IW. This study had two parts: (1) To test whether the weight-height cut-off values for weight deficit (%IW 90%, BMI 20 kg/m2) and obesity (%IW 120%, BMI 30 kg/m2) are compatible, we performed linear regression of BMI on %IW in peritoneal dialysis (PD) patients without amputations. In 349 men, BMI = 0.834 + 0.226 (%IW), r = 0.979. From this regression, the 95% confidence interval (CI) of BMI is 19.2-23.1 kg/m2 if %IW is 90%, and 26.1-29.9 kg/m2 if %IW is 120%. In 260 women, BMI = 2.194 + 0.184 (%IW), r = 0.974. From this regression, the 95% CI of BMI is 15.7-21.8 kg/m2 if %IW is 90%, and 21.3-27.3 kg/m2 if %IW is 120%. (2) To identify the direction and magnitude of the error of uncorrected BMI (BMIu) in dialysis patients with amputations, we analyzed weight-height indices in two groups of men by the computer nutrition program, which corrects %IW, but not BMI for amputation, and by the corrected BMI (BMIc) formula. In group A (amputation without height loss, n = 11), %IW = 110.2% +/- 16.9%, BMIu = 23.6 +/- 2.7 kg/m2, BMIc = 26.4 +/- 3.8 kg/m2 (p < 0.001, BMIc vs BMIu), and 5 of the 11 BMIu values fell below the 95% confidence band of the regression of BMI on %IW in patients without amputations. In group B (amputation with loss of height, n = 6), %IW = 92.7% +/- 19.9%, BMIu = 33.9 +/- 10.7 kg/m2, BMIc = 22.1 +/- 4.4 kg/m2 (p < 0.005, BMIc vs BMIu), and 5 of the 6 BMIu values fell above the 95% confidence band of the regression of BMI on %IW in patients without amputations.
CONCLUSIONS: (1) The weight deficit cut-offs for %IW and BMI are compatible in non amputated men and women. (2) The obesity cut-offs for %IW and BMI are compatible in non amputated men, but not in non amputated women. (3) Amputation without height loss decreases BMIu, while amputation with height loss increases, in general, BMIu. (4) BMI should be corrected in PD patients with amputations.