Objective. To investigate the articular toxicity of 2 aluminum derivat
ives, one insoluble (hydroxide) and/or the other soluble (lactate), af
ter a single administration in rabbits and rats. Methods. First, alumi
num levels in plasma, urine, synovial tissue, liver and kidney were me
asured in saline treated rabbits and I to 2 days after an articular in
jection of 75 mg of aluminum compounds into their right knee. The meth
odology used was argon plasma emission spectrometry. Thereafter, the j
oint toxicity of aluminum lactate at the same dose regimen was evaluat
ed for 2 days by a qualitative histological examination of synovial ti
ssue and articular surfaces and a colorimetric assay (1,9-DMB) of pate
llar articular cartilage proteoglycan content. Secondly, the single in
jection of 50 mg of aluminum derivatives as an inducer of inflammation
was studied in the rat subcutaneous air pouch, a model for a synovial
-like space. Leukocytes and eicosanoids levels were measured in pouch
washout fluids from 1 to 72 h after injection. Results. After injectio
n into rabbit knee, aluminum lactate largely distributed within the bo
dy while hydroxide remained locally. However, aluminum lactate resulte
d in perivascular edema, sparse infiltration of inflammatory cells in
the synovium and a hemorrhagic effusion. Proliferation of the synovial
cell layer coexisted with an apparent loss of proteoglycan in superfi
cial zones of tibial and femoral cartilages when patellar proteoglycan
content remained unchanged. Aluminum hydroxide did not affect joint s
tructures. In the air pouch experiment, aluminum lactate increased pro
staglandin E(2) (PGE(2)) levels from 3 to 10 h after its injection and
less intensively leukotriene B-4 (LTB(4)) levels after 6 h, in the ab
sence of leukocytes migration into the cavity. In contrast, aluminum h
ydroxide increased leukocytes count in pouch-washout fluid from 3 to 2
4 h after its injection when PGE(2) and LTB(4) levels were little modi
fied. Conclusion. Although some differences attributable to dissimilar
ities in the experimental model used, aluminum compounds, even in a so
luble form, may damage joint structures either directly or through sti
mulating the secretion of eicosanoids by synovial-like cells.