Initial safety data comes from the pre-approval clinical trials, as summarized by Male (2021):
Note that pregnant women were not enrolled in the trials, and participants were asked to avoid conception, so these numbers are presumably of accidental pregnancies. Just eyeballing the results, there is no obvious difference in fertility or miscarriage between the control and experimental groups.
Approval of mRNA vaccines for women of childbearing age was also influenced by the long safety track record of previous mRNA vaccines studied in humans since 2008, animal studies demonstrating no reproduction-related safety concerns, and the high risk to pregnant women from infection.
Several subsequent clinical trials demonstrated no difference in women's fertility pre and post vaccination (Safrai et al, 2021; Orvieto et al, 2021; Morris, 2021).
The pre-approval clinical trials are placebo-controlled RCTs, meaning that participants were randomly injected with a vaccine or saline. We now have far more real-world data, but such data is no longer placebo-controlled, randomized, blinded, or matched, so is in some ways lower quality, but this is substantially made up for by the much larger numbers and longer timeframe involved.
Shimabukuro et al (2021) first reported on 827 women in the US, vaccinated before or during pregnancy, who completed their pregnancy at follow-up:
As an example of reading this table, congenital anomalies have a background (ie, unvaccinated) rate of about 3%; vaccinated women in the study had a rate of 2.2%. All the measures reported compare favourably against the background incidence rate. You may notice the missing data for spontaneous abortions - this was addressed in a separate note confirming no difference (also see Kharbanda et al, 2021).
Unvaccinated women tend to be less educated, lower socioeconomic status, and may have predisposing health factors that compare unfavourably for reasons unrelated to the vaccine. A larger study by Goldshtein et al (2021) of 1387 pregnant women in Israel attempts to compensate for some of the drawbacks of real-world data by comparing against 1427 unvaccinated pregnancies during the same period, matched for demographics:
Again, vaccination does not appear to affect outcomes in any way for participants who completed pregnancy at follow-up.
A more recent study by Wainstock et al (2021) of 913 vaccinated pregnant women matched with 3486 unvaccinated controls gives similar results:
... no differences were found between the groups in pregnancy,
delivery and newborn complications, including gestational age at
delivery, incidence of small for gestational age and newborn
The results are too many to include here, but check out Table 3 for relative risk (odds ratios) of 15 outcome variables measured.
The number of women enrolled in the above studies is relatively small compared with the hundreds of thousands who have been vaccinated prior to or during pregnancy around the world. Outcomes for these women are tracked more broadly by various health agencies as part of adverse event reporting. These agencies continue to report no difference in the rate of adverse events related to pregnancy and childbirth compared to background (Canada, UK, USA).
The real-world comparison data consists of unvaccinated uninfected pregnancies. Unvaccinated infected pregnancies have substantially worse outcomes, as summarized in a meta-analysis by Allotey et al (2020):
For example, the ICU admission rate for infected women is more than 18 times higher than uninfected women, and the neonatal death rate is more than twice as high. All the comparison rates (odds ratios) are substantially worse for unvaccinated pregnant women who become infected during their pregnancy (also see Ko et al, 2021; Wei et al, 2021; Villar et al, 2021).
Approximately 10% of the 67,271 women included (worldwide) in the above meta-analysis contracted COVID-19 during their pregnancy, and there is clearly a substantial risk associated with infection that led to the prioritization of pregnant women for vaccination as a particularly vulnerable group. On the other hand, women vaccinated prior to or during pregnancy do not evidence any difference from uninfected women. In other words, not vaccinating carries a substantial risk from infection, while vaccination does not appear to add any risk for women of childbearing age.