Between December 1983 and August 1992, 21 knees in 19 patients with sy
mptomatic total knee replacements were arthroscopically diagnosed. The
average time between total knee replacement and arthroscopy was 20 mo
nths (ranging from 4 to 84 months). All patients were primary total kn
ee replacements. Metal bars, 5 and 8 mm in diameter and 40 cm in lengt
h, were made for arthroscopic treatment of fibroarthrosis. Thirteen kn
ees of 11 patients with a diagnosis of arthrofibrosis had an average i
mprovement of 42 degrees are of motion 1 year after arthroscopic surge
ry. Three knees with an average of 15 degrees increase of motion were
determined to be failures. Another two patients had arthroscopic resec
tion of fibrous bands with complete relief of patella pain. Among six
patients who had revision of total knee replacements after arthroscopi
c diagnosis, four had wear in the metal backed patella components, and
two had wear in the tibial insert and loosening of cementless patella
component. An early diagnosis of implant failure under arthroscopic c
ontrol was made, which easily allowed revision of the metal-backed pat
ellar button before the development of metallosis and massive osteolys
is caused by the marked wear of polyethylene. With the use of our spec
ially made metal bars for treating arthrofibrosis, we could release th
e adhesion more easily and avoid damage of valuable arthroscopic instr
uments.