Background The age-standardized incidence of primary liver cancer (PLC) has
apparently increased in most countries, The aim of this study was to estab
lish the degree of the stated increased incidence of PLC with a more precis
e model and to determine the role of age for incidence of PLC,
Methods Preserved biopsy specimens from all PLC cases in a well defined pop
ulation (Goteborg, Sweden) from a 22-year period, when the age-standardized
incidence rate of PLC was substantially increased, were reevaluated and cl
assified histopathologically. A Poisson model was used to calculate time tr
ends in incidence of the hepatocellular (HCC, n = 532) and cholangiocellula
r (CCC, n = 48) type of PLC and, to estimate the role of age for the risk o
f PLC, Risk coefficients for age, sex and calendar year mere estimated by d
ata from the material,
Results Adjusting for age there was no significant change in incidence of P
LC during the study period, The age risk coefficient was for HCC 0.102 in m
en and 0.082 in women, and for CCC 0.090 in men and 0.113 in women, These r
isk coefficients differed significantly (p < 0.001) from those concerning a
ll cancer in western Sweden with exception for CCC in men,
Conclusions The generally accepted age-standardized incidence method is not
fully accounting for the ageing effect in studies of PLC, which was found
to be an extremely age-dependent cancer, Use of this method instead of the
Poisson model may give an impression of a substantial increased incidence o
ver time.