Four cytology sampling methods were compared in 1063 patients referred
for colposcopy with a recent abnormal smear. A dyskaryotic smear of a
ny grade was considered a positive result, though comparisons were lim
ited to cases with a subsequent biopsy confirming CINII or III. There
were no differences between the abilities of any of the four methods t
o detect higher grades of CIN (chi2(3)=4.603, P>0.20). The presence or
absence of endocervical cells in a smear was not significantly associ
ated with any variation in success rate (chi2(1) =0.959, P>0.30). The
joint analysis of the four methods and the presence/absence of endocer
vical cells also showed no significant effects (chi2(7)=12.768, 0.1>P>
0.05). In the latter analysis the trend towards a conventional level o
f significance was accounted for by the Aylesbury spatula giving a rel
atively high success rate when endocervical cells were present. The su
ggestion of advantage for the Aylesbury spatula merits further investi
gation.