Standardisation of uterine natural killer (uNK) cell measurements in the endometrium of women with recurrent reproductive failure

- There is variation in the numbers of endometrial uNK cell reported by different centres.
- We have identified the factors which contribute to these differences.
- Adoption of a standard protocol reduced the variation considerably.
- Use of a single method is needed for measurements to have clinical significance.

Considerable work is being carried out on endometrial NK cells to determine whether they play a role in successful pregnancy outcome. In addition there is debate about whether measurements of uNK should be included in the clinical assessment for women with recurrent implantation failure or recurrent miscarriage. A hindrance to taking this forward is the fact that the density of uNK cells reported by different centres is very different. The aim of this study was to determine the reason for these differences and to develop a standardised method. Three centres participated in the study. Each centre exchanged five formalin fixed, wax embedded sections of endometrium from five women. Sections were immunostained for CD56. Images were taken of 10 random fields at ×400 magnification; total stromal and uNK cells were counted using Image J. Results were expressed as % positive uNK cells and the variation in counts obtained in each centre was compared. After initial analysis a standardised protocol was agreed and the process repeated.Significant variation was seen in the counts obtained after initial analysis (Centre A vs.B, mean difference = −0.72 P < 0.001; A vs.C mean difference = −0.47 P < 0.001; B vs.C, mean difference = 0.25 P = 0.085). Analysis suggested that differences may be due to duration of tissue fixation, the embedding and sectioning processes, selection of areas for assessment, definition of immunopositive cells and inclusion or exclusion of blood vessels. Adoption of a standardised protocol reduced the variation (Centre A vs.B mean difference = −0.105 P = 0.744; A vs.C mean difference = 0.219 P = 0.150; B vs.C mean difference = 0.32 P = 0.031). Use of a standardised method is needed to establish a normal range for uNK cells and to develop a meaningful clinical test for uNK cell measurements.