Predictive value of symptom level measurements for complex regional pain syndrome type I

The validity with respect to presence or absence of CRPS I according to Veldman's criteria was assessed for measured pain, temperature, volume differences and limitations in range of motion. Evaluated were 155 assessments of 66 outpatients, initially diagnosed with CRPS I, but many of them not so on follow up visits. Pain was measured with VAS and McGill, temperature by infrared thermometry, volume differences by water displacement volumeters and limitations in range of motion by universal goniometers. Sensitivity, specificity, positive and negative predictive value of the measurement instruments at different cut-off points was calculated. Combined symptom scores were evaluated in a similar fashion. High sensitivity was found for the VAS, McGill, and range of motion. The specificity was overall lower, but highest values were obtained for volume differences. The positive predictive value was good for all measurement instruments. Negative predictive value was lower, especially for measurement of temperature and volume asymmetries. If sensitivity and specificity are equally important, VAS>3 cm, McGill>6 words, temperature difference>or=0.4 degrees C, volume difference>6.5% and ROM limitation>15% provide the best results. Using these cut off values, the highest value of sensitivity and of sensitivity and specificity combined, was found for a combination of VAS, McGill and ROM. The highest value of specificity was found for different combinations of 3, 4 and 5 instruments, all containing the VAS. We conclude that the measured pain, temperature, volume and range of motion can be used as diagnostic indicators for establishing presence or absence of CRPS I.