e-mail: gsacks{at}hhnt.nhs.uk
Sir,
We read with interest the recent article by Almagor et al. (2004) investigating changes in C-reactive protein (CRP) levels in women treated by IVF. We have also investigated CRP levels in such women, and our conclusions, soon to be published in this journal (Sacks et al., 2004
), appear to contradict those of Almagor et al. It was interesting that they resorted to using a hitherto unused measure of CRP (the ratio of CRP levels on days of embryo transfer and oocyte retrieval) to report a statistical difference between women who became pregnant and those who did not. In fact, there was no significant difference in direct comparisons at any time between oocyte retrieval and day 12. On closer inspection, however, women who became pregnant had higher CRP levels at every time of sampling, which is consistent with our findings. There are a number of reasons why Almagors data may not have similarly reached statistical significance. First, the CRP assay used had a lower limit of detectability of only 5 mg/l (high sensitivity assays with limits of 2 mg/l are available). Small differences will be detected more reliably with the high sensitivity assay. Secondly, bias cannot be excluded as they failed to include all women treated. Indeed, 20 cases (nearly 20% of the treated cycles) had incomplete data and were excluded from analysis. These factors, combined with the relatively small numbers in the study, significantly undermine the strength of the study and may explain the failure to reach statistical significance in the pregnant versus non-pregnant comparisons.
We would also like to offer an explanation for the difference that is reported: that the day of transfer/retrieval CRP ratio was lower in women who became pregnant. It is now well recognized that ovulation is an inflammatory process, and surgical oocyte collection is likely also to increase inflammatory markers. Thus one would suspect that women with more oocytes collected would also have higher CRP levels around the time of oocyte collection. It is also well documented that pregnancy rates are related to the number of oocytes collected at IVF. Thus women who fell into the pregnant group would be predicted to have higher CRP levels at oocyte collection, as appears to be the case. Unfortunately, data on the number of oocytes collected were not presented. We are therefore not surprised that the relative fall in CRP levels from oocyte collection to day of embryo transfer was greater for women who became pregnant. This does not mean that pregnancy itself is associated with lower CRP levels. On the contrary, our data, soon to be published in this journal, showed that pregnancy after IVF treatment is associated with raised CRP levels on day 14 after oocyte collection. The real conclusion from Almagors report is consistent with that, with pregnancy associated with higher mean CRP levels on both days 57 and day 12.
References
Almagor M, Hazav A, Yaffe H (2004) The levels of C-reactive protein in women treated by IVF. Hum Reprod 19, 104106
Sacks GP, Seyani l, Lavery S and Trew G (2004) Maternal C-reative protein levels are raised at 4 weeks gestation. Hum Reprod 19, in press.