SARS-CoV-2 variant-related abnormalities detected by prenatal MRI: direct or indirect effects of the infection must be demonstrated—Authors’ reply

We read the Correspondence from Dr. Josef Finsterer¹ regarding our manuscript on SARS-CoV-2 variant-related abnormalities detected by prenatal MRI² with great interest. Dr. Finsterer's main concern about our work was the evidence of placental or fetal infection with the virus. We would like to reply to his comments.

SARS-CoV-2 infection during pregnancy raises several important questions, which are far-reaching and beyond the expertise of a single, specific medical field. We would like to emphasize that our study is the result of a multidisciplinary collaboration between (neuro)radiologists with long-standing expertise in fetal MR imaging, gynecologists, pathologists, virologists, and internal medicine specialists with expertise in coagulopathies.

Currently, there is undisputable evidence of vascular placental impairment in COVID-19, as shown by previous studies that have examined placental histology.^3,4 The primary aim of our study was to report the frequency and visualize the MR pattern of placental abnormalities after SARS-CoV-2 infection during pregnancy—as demonstrated by fetal in vivo imaging. Furthermore, we were able to support our statistically significant findings of specific MR changes of the placenta (detection of globular-shaped placentas, increased frequency of placental lobules and hemorrhages, change in the extent of placental lobulation and hemorrhages, increase in the thickness of the placenta, and an increase in the frequency of FGR) by histological analysis in 12 cases. We agree with Dr. Finsterer that immunohistochemistry of the placenta may have provided further interesting insights, but would not have changed our results in any way. As shown previously,5, 6, 7 the rate of vertical virus transmission is only 0–3.2%; thus, we would not expect informative results from either monitoring the infection status by performing RT-PCRs on fetuses in vivo and in utero (strongly opposing the noninvasive character of our noninvasive imaging approach), or by acquiring information from postnatal PCRs in neonates several months after maternal infection.

Here, we would like to reiterate that we have concluded from the results of our study, in conjunction with the existing literature, that the placenta serves as a sufficient barrier against the virus. However, the placenta may undergo variant-specific damage from SARS-CoV-2 and fetuses may subsequently suffer harm from placental malperfusion. The vertical virus transmission rate of 0–3.2% cannot solely be blamed for an FGR rate of up to 25%.

Of course, it is possible that placentas that were not exposed to the virus may also develop changes at later gestational weeks for reasons other than SARS-CoV-2 infection.

However, we would like to point out that we mentioned this clearly in our publication and that this is exactly the reason we conducted a case–control study with appropriate statistics. Placental changes were also partially evident in the healthy, matched control cases, but these changes were, by far, not as frequent or drastic in extent as after SARS-CoV-2 infections.

We fully agree with Dr. Finsterer that factors that could have an impact on the placenta or the fetus are critical points in prenatal diagnostic case–control studies. Therefore, we addressed this in the Limitations and mentioned there that we did not include any case with a disease or obstetric event that could have had an impact on the fetus or placenta.

None of the pregnant individuals (neither in the infected nor in the control group) had gestational diabetes. Preeclampsia occurred later in gestation in one pregnancy, as described in the Fetal Changes subsection of the Results.

We want to be clear that we, indeed, paid very close attention to other potentially infectious diseases, and therefore, also excluded one case with CMV infection during pregnancy and mentioned this in the Study Population section. Moreover, we would like to mention that the requested information on the vaccination status of the pregnant women (vaccinated or unvaccinated, number of doses, placental and fetal outcome) is described in detail in the Results section and discussed in the Discussion section. Furthermore, we want to address the requested matching by maternal medication and emphasize that, based on the legal situation and the ethical and medical standards in Austria, of course, no patient received medication during pregnancy that could have caused damage to the placenta or fetus and that a statistical bias could not have occurred in this regard.

Regarding variant-specific outcomes, we would like to point out that histopathological data already exists, with high case numbers, which have investigated placental tissue and which support the extent of lesion-specific variant-related changes.⁸ In our prenatal MRI study, the variant-specific changes found were statistically significant, despite the limited sample size.

We would like to clarify that, according to the recommendations of the International Society in Ultrasound and Obstetrics,⁹ blood-sensitive MR sequences in fetal MRI should be T2∗ sequences, as the implementation of SWI is rather recent in the field of fetal MRI (see also detailed acquisition parameters and protocol in Table 3 in the Supplementary Material of the paper). Based on our previous experience and comparative studies on MR imaging findings and placental histology,¹⁰ we were able to show that MRI changes do accurately reflect placental histopathology.

Based on our expertise in MR imaging, there is no doubt that the lesion in Figure 3 is a hemorrhage and that this was confirmed on follow-up MRIs, as well as by postnatal clinical manifestation. We can definitely exclude the suggested differential diagnoses, such as turbulent or rapid CSF flow, or aerated spaces. Gadolinium-based effects can be generally excluded as—opposed to adult neuroimaging—in fetal MRI, gadolinium is generally avoided and was not used in any of our cases (also see ISUOG Practice Guidelines⁹).

Finally, we agree with Dr. Finsterer that research on the health effects of SARS-CoV-2 harbors the risk of drawing scientifically unsupported conclusions.

However, we hope that informed readers will agree that our clarifications, the data presented in our paper, and our thorough multidisciplinary approach speak for themselves. These prenatal MRI data indicate a strong association of SARS-CoV-2 infection and placental abnormalities, with the potential for fetal impairment.

Together, we have given extensive thought to minimizing potential effects that could bias our results; however, we are, of course, open to criticism or comments on our study.

Contributors

All authors contributed to the conception and writing of the Reply.

Declaration of interests

The authors declare no conflict of interest.