The purpose of this article is to evaluate arm measurements of breast cancer patients to critically assess absolute change in arm size compared to relative arm volume change as criteria for quantifying breast cancer-related lymphedema (BCRL). We used pre-operative measurements of 677 patients screened for BCRL before and following treatment of unilateral breast cancer at Massachusetts General Hospital between 2005 and 2008 to model the effect of an absolute change in arm size of 200 mL or 2 cm compared to relative arm volume change. We also used sequential measurements to analyze temporal variation in unaffected arm volume. Pre-operative arm volumes ranged from 1,270 to 6,873 mL and correlated strongly (Kendall’s τ = 0.55) with body mass index (BMI). An absolute arm volume change of 200 mL corresponded to relative arm volume changes ranging from 2.9 to 15.7 %. In a subset of 45 patients, modeling of a 2-cm change in arm circumference predicted relative arm volume changes ranging from 6.0 to 9.8 %. Sequential measurements of 124 patients with > 6 measurements demonstrated remarkable temporal variation in unaffected arm volume (median within-patient change 10.5 %). The magnitude of such fluctuations correlated (τ = 0.36, P < 0.0001) with pre-operative arm volume, patient weight, and BMI when quantified as absolute volume change, but was independent of these variables when quantified as relative arm volume change (P > .05). Absolute changes in arm size used as criteria for BCRL are correlated with pre-operative and temporal changes in body size. Therefore, utilization of absolute volume or circumference change in clinical trials is flawed because specificity depends strongly on patient body size. Relative arm volume change is independent of body size and should thus be used as the standard criterion for diagnosis of BCRL.