Meta-analysis: Increased Risk for Birth Defects from IVF/ICSI vs Spontaneous Conception

April 23, 2012Carole 2 Comments »

The on-line version of Fertility and Sterility published a review of 56 previous studies; “Birth defects in children Conceived by in vitro fertilization and intracytoplasmic sperm injection: a meta-analysis”on April 3, 2012, epub ahead of print. This study found an increased relative risk (RR) of congenital birth defects reported in children conceived with either IVF or ICSI compared to spontaneously conceived children. The study found no difference between the relative risk of birth defects between IVF and ICSI, but both IVF and ICSI techniques showed elevated birth defect risk compared to spontaneously conceived children.

Relative risk is a statistical term used to express the relative probabilities of an event occurring in two groups, usually a control and experimental group. If the relative risk is 1.0, than the probability of an event occurring is the same in both groups. If the RR is greater than 1.0 (eg, 1.37) as found in this study, the risk is elevated in one group relative to the other. What this means is that the probability of having a child with birth defects is elevated if that child was conceived using IVF or ICSI compared to spontaneous conception. Specifically, compared to spontaneous conception, use of ART increases the risk of birth defects by 37% over spontaneous conception. So if the probability of birth defects is (for example ) 3% in the spontaneously conceived population, the probability of birth defects in the IVF/ICSI population would be a little over one-third (37%) more or about 4%. It does not mean that the risk of birth defects is 37% if you use IVF or ICSI to conceive your child. More simplified explanations of odds ratios and risk ratios can be found here.

I am not a statistician but I suspect statisticians may take issue with this paper and argue that odds ratio might have been a better method to compare the two groups or that some studies were improperly excluded or included, but I can’t address any of that. What is sobering is that even if you quibble over the best way to have performed the meta-analysis, it still doesn’t fundamentally change the take-home message which is that in a review of 46 studies involving 124,465 children comparing the mechanisms of conception, most of the studies (n=38) showed elevated risk with the relative risk ratios varying from 1.03 to 5.53.

So what could this mean if the analysis is correct? There are at least three proposed explanations for this finding:

  • ¬†ART procedures themselves (ovarian stimulation or technical lab interventions) somehow increase the risk of birth defects.
  • The reported rates for birth defects are higher in the ART population than the spontaneously conceived population because ART children are studied more closely from conception, thus falsely raising the apparent relative risk.
  • There are more underlying genetic conditions in the ART conceived population than the spontaneously conceived population and these genetic conditions are revealed through birth defects when otherwise infertile patients conceive using ART. If this is true, the proper control group would be subfertile patients who conceived spontaneously after some time compared to subfertile patients who conceived using ART. Unfortunately birth outcome data on spontaneously conceiving subfertile population is not readily available.

Already, bioethicist Art Caplan has chimed in with his usual dire warnings about IVF , “Time to think of health costs to IVF babies, bioethicist says” in which by not explaining relative risk and characterizing this increased risk as “huge”, he implies that the 37% increased risk is an absolute, not increased relative risk, implying that using IVF gives you a greater than one-in-three chance of having a child afflicted with a birth defect.¬† While I disagree with Dr. Caplan’s sloppy use of statistics, I agree wholeheartedly¬† that we should all be cognizant of the possible implications of these findings and more long-term studies should be performed. I would even go further and suggest that professionals in the ART field should clamor for federal funding of longitudinal studies of children born using ART so that we can finally settle this question and be able to better advise patients of their actual risks depending on which scenario above (or alternative explanation) best explains these findings.






© 2012, Carole. All rights reserved.

2 Responses to this entry

  • It Is What It Is Says:

    Hmmm, this is interesting insofar as they have long been trying to determine whether this is a causal link between ICSI specifically and IVF more broadly and any sort of birth defect. Did the study mention the types of birth defects?

    We did old-fashioned PGD with our successful cycle with our now 5 year old son and early on I was worried about whether the procedure itself and the extraction of the cell for biopsy would pose any risks to him. At this early age, it doesn’t appear so (in that he was not born with any congenital defects) and I hope that remains so down the line.

  • Carole Says:

    Hello “It is what it is,
    Thanks for your comment and question. Regarding types of defects: The only specific remark about the types of defects was that “effect of ART on the nervous system is relatively obvious compared to effects on eyes, ears, face and neck, which may suggest that earlier developed systems were more sensitive to birth defects by ART”. In another section, they mention that musculoskelatal birth defects did not reach statistical significance.

    Re PGD; I wouldn’t worry about the biopsy for PGD causing birth defects. On day 3, biopsy of one cell of 8 totipotent cells is removing one clone cell from a group of 8 identical clone cells (assuming no mosacism) so the remaining cells should not have lost any important genetic information. The problem that might arise is that the embryo does not progress, not that limbs (for example) are missing due to loss of one cell. On day 5, the trophectoderm of the blastocysts is sampled which does not contribute any cells to the future fetus so again, should not be able to cause fetal birth defects.

    I would ask your pediatrician to be sure but I suspect if your son hasn’t shown any issues by age 5, he should be fine. Best wishes, Carole

Join the discussion