Digital and manual subtraction images obtained during the arthrographic evaluation of 78 painful hip prostheses were reviewed retrospectively. Revision arthroplasty was performed in 53 of these cases, and the arthrographic and surgical findings were correlated. The digital and manual subtraction images were evaluated without knowledge of the surgical results using established criteria for component loosening. Digital subtraction arthrography of the femoral component demonstrated a 96% sensitivity and 100% specificity for the diagnosis of component loosening. Sensitivity and specificity for acetabular component loosening were 83% and 80%, respectively. Plain film subtraction of the femoral component demonstrated a 79% sensitivity and 100% specificity; the sensitivity and specificity for the acetabular component were 75% and 80%, respectively. The difference between detection of femoral component loosening on digital as opposed to manual subtraction images was statistically significant (P<0.05). this="" study="" demonstrates="" that="" digital="" subtraction="" improves="" the="" evaluation="" of="" femoral="" component="" loosening="" in="" painful="" hip="">