Objective: Data about the diagnostic efficiency of the 99mTc-MDP triple-phase bone scintigraphy used in identification of peri-prosthetic infection showed a remarkable variability because of using different scintigraphic interpretation criteria.The purpose of this study is to investigate the uptake features of the normal, aseptic loosening and infected artificial hips and knees in the blood pool phase, in patients with joint pain after arthroplasty.Methods: During February 2008 to June 2010, 113 consecutive patients (35 men, 78 women, with mean age of 65.7±10.4 yr) with joint pain after arthroplasty were studied retrospectively.In these patients, total of 154 joints were replaced (72 hips, 82 knees), and 116 of them (63 hips, 53 knees)were symptomatic.Except of blood perfusion and bone imaging, blood pool imaging was performed at 5-10min after tracer injection in each patient, and the presence of infection was considered when abnormal soft tissue uptake surrounding the prosthesis was found.Clinical diagnosis of peri-prosthetic infection was based on the visibly purulent synovial fluid, positive culture obtained by joint aspiration or surgery, acute inflammation confirmed by histopathologic findings, or an over six months' followed up study.The diagnostic efficiency of blood pool scan in identification of peri-prosthetic infection was evaluated for the hips, knees and overall,respectively, by comparing with the final clinical diagnosis.As a semi-quantitive criterion,peri-prosthetic soft tissue uptake to mid-thigh soft tissue uptake ratio was measured for the un-replaced joints (52 hip, 24 knee), as well as the normal, aseptic loosening and infected artificial joints.Results: Within the 116 symptomatic joints, infection was clinically diagnosed in 59 (19 hip, 40 knee), and the other 57 (44 hip, 13 knee) was aseptic loosening.Table 1 shows the diagnostic efficiency of blood pool imaging used in identification of infected prosthesis.The diagnostic specificity was significantly different between the hip and knee (x2=12.263, P<0.001).States of the peri-prosthetic soft tissue uptake in the each kind of joints were shown in Table 2, and it was suggest that the soft tissue uptake of infected joints is obviously higher than that of the aseptic loosening and normal artificial joints, for both hip and knee.However, there was no significant different in the soft tissue uptake between aseptic loosening hips and normal artificial hips,whereas the soft tissue uptake in aseptic loosening knees was significantly higher than that of the normal artificial knees (t=2.535, P=0.015).In addition, comparing with the hips, soft tissue uptake was lower in those un-replaced and normal artificial knees, but higher in those infected artificial knees.Conclusions: Blood pool imaging in 99mTc-MDP triple-phase bone scintigraphy is useful in identification of peri-prosthetic infection in patients with painful joint after arthroplasty.Because of the variance of peri-prosthetic uptake between artificial knee and hip joint, differential diagnostic criteria should be further invested.