Abstract
The appearance of Grannum Grade (GG) 3 changes in the placenta around 34–36 weeks is a predictor of adverse perinatal outcome which may be reduced by reporting GG to the clinician. This has led to the suggestion GG to be reported during any third trimester scan. There are no published data on the reproducibility of GG. We therefore set out to evaluate intra- and interobserver variation in GG. Fifty-five placental images from normal and complicated pregnancies of different gestational age were collected between April and October 2001. Three fetal medicine consultants and three experienced ultrasonographers graded the images as 0, 1, 2, 3 or ungradeable; 4–6 weeks later they regraded the same images, presented in a different order and with different codes. Observers were blinded to their previous grading and to each other. All the data were analysed using the package Stata version 7.0 (Stata Corporation, Texas). Weighted kappa (κ), with linear weights, was used to look for strength of agreement. There was good agreement between the two observations of each placental image for five observers (κ 0.61–0.89) and moderate agreement for one (κ 0.55). However, the κ values for comparisons between the 15 pairs of observer ranged from 0.23 to 0.69 with six values below 0.41, indicating only fair agreement. This is confirmed by the overall κ value of 0.24 between all six observers.