Changes in circulating levels of catecholamines, cortisol, glucose, interle
ukin-6 and C-reactive protein and in the leucocyte count were investigated
for 7 days after surgery in 158 patients undergoing hip or knee arthroplast
y. We compared the responses to the two operations, and also examined the e
ffects of pathology (osteoarthritis and rheumatoid arthritis) on the change
s associated with knee arthroplasty. Exploratory factor analysis was applie
d to the data to identify the variables and sampling times that could be us
ed in future to provide a concise description of the response. Patients und
ergoing knee arthroplasty showed significantly greater changes in noradrena
line, adrenaline and glucose levels. but not in cortisol levels, compared w
ith those undergoing hip arthroplasty. Interleukin-6 and C-reactive protein
concentrations were also significantly greater in knee patients than hip p
atients; however, when corrected for pathology, many of these differences w
ere not significant. Minimal effects of pathology (chronic inflammation wit
h rheumatoid arthritis) were found on the hormonal changes in knee patients
. In particular, there was little evidence to support the inference from an
imal data that the hypothalamic-pituitary-adrenal axis is impaired. The exp
ected increases in interleukin-6 and C-reactive protein concentrations were
found in the rheumatoid arthritis patients. Exploratory factor analysis sh
owed that the response could be separated into six components, accounting f
or 60 % of the total variance, and identified the variables and sampling ti
mes indicative of each. In conclusion, there are differences in the hormona
l, but not inflammatory, responses to hip and knee arthroplasty. Little evi
dence was found for an important effect of pathology on the changes associa
ted with knee surgery. Factor analysis provided a useful summary of the dat
a.