BACKGROUND: Limb circumference measurements (CM) and perometry are the preferred methods for objectively measuring arm volume in lymphedema surgery research. Understanding the measurement bias involved in these measuring systems is important to properly interpret and compare studies and their results. METHODS: Arm volumes from 91 patients were measured using sequential girths and the truncated cone formula (CM) and with the use of an automated perometer (perometry). The absolute volume of the largest arm (V), the volume difference between the arms (VD), and the relative difference between them (percentage of excess volume [PEV]) were calculated with both methods. The agreement between methods was assessed by the Pearson's correlation test and the Bland-Altman's method. RESULTS: Correlations were strong for V (r = 0.99), VD (r = 88), and PEV (r = 0.86). Volumes measured by perometry were, on average, 10.6 mL smaller than volumes calculated from CM, while their limits of agreement (LOA) ranged from -202 to 181 mL. The LOA represents the range we could expect the arm volumes measured with the two methods to differ by chance alone, 95% of the times. For VD, LOA was -101 to 141 mL, with a mean difference of 19.9 mL, while PEV had a mean difference of 0.9%, with LOA ranging from -5 to 6.8%. CONCLUSION: There is considerable measurement error between arm volume estimated by perometry and by CM. Volumes calculated with these methods should be compared with caution. Furthermore, we observed an increasingly relevant measurement bias in outcomes that are mathematically derived from arm volumes. Thieme Medical Publishers 333 Seventh Avenue, New York, NY 10001, USA.
BACKGROUND: Limb circumference measurements (CM) and perometry are the preferred methods for objectively measuring arm volume in lymphedema surgery research. Understanding the measurement bias involved in these measuring systems is important to properly interpret and compare studies and their results. METHODS: Arm volumes from 91 patients were measured using sequential girths and the truncated cone formula (CM) and with the use of an automated perometer (perometry). The absolute volume of the largest arm (V), the volume difference between the arms (VD), and the relative difference between them (percentage of excess volume [PEV]) were calculated with both methods. The agreement between methods was assessed by the Pearson's correlation test and the Bland-Altman's method. RESULTS: Correlations were strong for V (r = 0.99), VD (r = 88), and PEV (r = 0.86). Volumes measured by perometry were, on average, 10.6 mL smaller than volumes calculated from CM, while their limits of agreement (LOA) ranged from -202 to 181 mL. The LOA represents the range we could expect the arm volumes measured with the two methods to differ by chance alone, 95% of the times. For VD, LOA was -101 to 141 mL, with a mean difference of 19.9 mL, while PEV had a mean difference of 0.9%, with LOA ranging from -5 to 6.8%. CONCLUSION: There is considerable measurement error between arm volume estimated by perometry and by CM. Volumes calculated with these methods should be compared with caution. Furthermore, we observed an increasingly relevant measurement bias in outcomes that are mathematically derived from arm volumes. Thieme Medical Publishers 333 Seventh Avenue, New York, NY 10001, USA.