Aim: Endpoints that are more common than revision would be more valuable in survival analysis. Assessment of unicompartmental knee replacement (UKR) by patientñs own evaluation may be sensitive as for total knee replacement (TKR). Cohort of patients treated by UKR has been studied for satisfaction and pain as endpoints and related to TKR. Materials: At the Centro di Chirurgia Protesica, Istituto Ortopedico ÒGaleazziÒ, Milano (I), more than 600 hip and knee replacements are performed per year. Among these, from February to April 2001, 109 patients have been treated for primary unilateral arthritis of the knee by the same surgeon, same OR team, and environmental conditions. 53 patients received unicompartmental knee replacement (Allegretto TM, Centerpulse) because of unicompartmental knee degeneration. 56 patients received total knee replacement (Innex TM, Centerpulse) for tri-compartmental arthritis. Patients have been preoperative evaluated for age, ASA score, KSS and Function. Postoperative recovery and physio funtions. At 18 months follow-up, cohort of UKR has been evaluated for satisfaction and pain as endpoints by telephone interview. The same has been for TKR group. Results: No differences were between the UKR and TKR cohorts for age at surgery (UKR: 68yrs, 86-45; TKR: 69yrs, 83-51); ASA score, KKS (UKR: 54,2pts, 76-25; TKR: 43,3pts, 75-10); function (UKR: 67,5pts, 95-20; TKR: 60,6pts, 95-10). No differences were for recovery period and rehabilitation program. Within the TKR group had 1 died for post op cardiopulmonary complication. At follow-up patients was judged by the author not eligible for any revision, on clinical and radiological examination. None was lost at follow-up, none revised. At 18 months follow-up, UKR cohort, 52 on 53 patients were satisþed by their knee, event rate 98% (89–100). 18 on 52 were in mild pain and drug use, event rate 34% (28–39), 4 of those in trouble with stairs. 1 patient was not satisþed. For TKR cohort, 49 on 55 patients were satisþed, event rate 89% (81–97). 24 on 49 reported pain and drug use, event rate 49% (42–56), 7 of those in stairs climb limits. 6 patients were not satisþed. Conclusion: Because of long time is needed to produce adequate number in joint replacement groups, statistical method relies on survivorship analysis for studying adverse events. However confounding factors, lost of follow-up, and revision as endpoint may mislead results (Murray DW. et al. JBJS 1993). Meanwhile, randomised clinical trial provides a cornerstone not only for evidence-based medicine but also for evidence public health, evidence based hospital administration, evidence based purchasing, and evidence based consumerism (Sackett DL et al. BMJ 1994).

Our TKR cohort pain endpoint survival rate is comparable with others in literature (Murray DW, Frost SJD. JBJS 1998). Comparing TKR with UKR groups has limitation and EBM is not eligible for such groups. However, cohort study for joint replacement, performed within a short period of time, and less confounding factors, increases assessment reliability and endpoints deþ nition. Overall, our result shows lower pain and higher satisfaction events rate for UKR than TKR.