Complex Regional Pain Syndrome type I (CRPS I) is an illness which usually occurs due to major or minor tissue injury to the extremities. Because a unique pathophysiological mechanism for CRPS I has not yet been established, the diagnosis is based on observation and measurement of clinical symptoms and signs. In this study, a comparison was made between three sets of diagnostic criteria (the IASP, Bruehl et al. and Veldman et al.) based on patient reports and physicians' assessments of signs and symptoms associated with CRPS I, in 372 outpatients suspected of having CRPS I. Agreement between CRPS I diagnosis among the three sets was poor (κ-range: 0.29-0.42), leading to positive CRPS I diagnoses according to Veldman et al.'s criteria in 218 cases (59%), according to the IASP in 268 cases (72%), and according to Bruehl et al. in 129 cases (35%). Significant differences in patient profiles were found between the diagnostic sets for the number of patients reporting continuing disproportionate pain, larger area affected than the initial trauma (both p < 0.001), increase of symptoms due to exercise (p = 0.009), edema (p = 0.015), temperature asymmetry (p = 0.015), hyperesthesia, allodynia (both p < 0.001) and hyperalgesia (p = 0.036). Similarly, significant differences emerged for physicians' observations of hyperesthesia and allodynia (both p < 0.001). Highest combined values of sensitivity (SE) and specificity (SP) for the strongest cases of presence (n = 108) or absence (n = 62) of CRPS I were found for reported hyperesthesia (SE + SP:165%), allodynia (160%), observed color asymmetry (162%), hyperesthesia (157%), temperature asymmetry (154%) and edema (152%). The lack of agreement between the different diagnostic sets for CRPS I and the different clinical profiles that result from it may lead to different therapeutic and study populations, hampering adequate treatment and scientific development for this illness. We propose explicit reference to diagnostic criteria used in studies, and registration in trials of a broad variety of CRPS I features, as used in this study, to make subgroup phenotyping and post hoc analyses based on different diagnostic criteria possible.