Case 1

This was a nulliparous 29-year-old woman with an uneventful clinical history except for slight overweight. She presented clinical features suggestive of COVID-19 and a chest computerized tomography scan revealed moderate pneumonia (10%–25% of lung volume) at 32 weeks of gestation. Complete prenatal steroid prophylaxis was given, and at 33 weeks of gestation, cesarean delivery was performed for category II nonreassuring cardiotocography (CTG). A female neonate (birthweight of 2130 g) was delivered and needed invasive ventilation and oxygen supplementation for mild perinatal asphyxia (Apgar score of 2 and 5 at 1 minute and 5 minutes, respectively; pH, 7.21; lactate, 6.9 mmol/L). During neonatal intensive care unit (NICU) admission, the Clinical Risk Index for Babies (CRIB-II) score was 4, and respiratory distress syndrome (RDS) was diagnosed, as the neonate had an oxygenation index of 7.9 and a lung ultrasound (LUS) score of 12. Poractant alfa (200 mg/kg) was administered, and a prompt response allowed extubation (at approximately 24 hours of life) on continuous positive airway pressure (CPAP), which was discontinued after 7 days.

Case 2

This was a 30-year-old nulliparous woman with early mild gestational diabetes mellitus (type A1). At 31 weeks of gestation, she presented with fever and laboratory features of COVID-19 after having been in contact with a SARS-CoV-2–positive family member. Complete prenatal steroid prophylaxis was given, and at 31 weeks of gestation, cesarean delivery was performed because of nonreassuring CTG (category II). A male neonate (birthweight of 1600 g) was delivered with good perinatal transition (Apgar score of 8 and 10 at 1 minute and 5 minutes, respectively; pH, 7.28; lactate, 2 mmol/L). The CRIB-II and LUS scores were 3 and 9, respectively, mild RDS was diagnosed, and the neonate was supported with nasal mask-delivered CPAP for 7 days.

Case 3

A 32-year-old (gravida 3, para 2) woman was admitted to the hospital at 29 weeks of gestation with clinical manifestations of COVID-19. After 24 hours, disseminated intravascular coagulation was evident, and CTG was abnormal (category III), so an emergency cesarean delivery was performed. A female neonate was delivered (birthweight of 1220 g) with satisfactory perinatal adaptation (Apgar score of 3 and 7 at 1 minute and 5 minutes, respectively; pH, 7.17; lactate, 7.1 mmol/L). CRIB-II was 8, RDS was diagnosed, and the neonate was intubated, received 200 mg/kg of poractant alfa, and, at 24 hours of life, was extubated on CPAP, which was discontinued after 1 week.

Case 4

This was a 30-year-old nullipara admitted to the hospital at 40 weeks of gestation while presenting mild symptoms of COVID-19. She had a fever during labor, and CTG was nonreassuring (category II); therefore, emergency cesarean delivery was performed. A male neonate was delivered (birthweight of 3640 g) with good perinatal adaptation (Apgar score of 10 and 10 at 1 minute and 5 minutes, respectively; pH, 7.24; lactate, 3.4 mmol/L). At 16 hours of life, the neonate presented with transient tachypnea and required CPAP for 3 days; in addition, early-onset sepsis was suspected but not confirmed.

Case 5

This represented the first case of ascertained transplacental transmission of SARS-CoV-2 and was previously described by Vivanti and colleagues. In this case, an abnormal category III fetal heart rate tracing was observed.

Case 6

This was a fetal demise occurring at 32 3/7 weeks of gestation in a male fetus, weighing 2248 g. An autopsy revealed no congenital malformation. COVID-19 was diagnosed in the mother, a 27-year-old (gravida 2, para 1) woman, 10 days before the stillbirth, which occurred by vaginal delivery. Increased uterine arterial resistance was noticed at 23 weeks of gestation and was the only anomaly during an otherwise uneventful pregnancy.

Virology

Real-time polymerase chain reaction (RT-PCR) was positive in placental tissue samples for all 6 cases. Moreover, amniotic fluid collected during cesarean delivery was positive for cases 1 and 2, whereas gastric aspirate was positive for case 3. Nasopharyngeal aspirates obtained during the first 48 hours of life were positive for all neonates (and at 7 days, for cases 3 and 4). Moreover, the first case had a positive RT-PCR in neonatal bronchoalveolar lavage fluid obtained before surfactant administration; the second case also had positive RT-PCR in neonatal stool on the second day of life; the fourth neonate also had positive RT-PCR in gastric aspirate, urine, and stool at days 1 and 7. The neonate of case 5 had a positive RT-PCR in multiple samples as described elsewhere (case 3 of the main text). Finally, the fetus of case 6 had a positive RT-PCR in lung tissue.

Perinatal outcomes

Case 6 resulted in fetal demise, whereas the other 5 neonates survived. Perinatal outcomes of the 5 neonates with transplacentally transmitted infections were compared with those of 14 neonates born from women with COVID-19 during the third trimester of pregnancy (ie, those enrolled in the C+P− and C+P+ groups for the translational study; provided below) without transplacental SARS-CoV-2 infection.

The clinical history of these cases is resumed as shown in Supplemental Table 1 , and an illustrative CTG from one of these cases is shown in Supplemental Figure 1 .

Viral receptors and viral load

Angiotensin-converting enzyme 2 (ACE2) receptor (overall P =.335) and transmembrane serine protease 2 (TMPRSS2) expression (overall P =.163) were similar in the 3 groups of placentas ( Supplemental Figure 2 ). The estimated placental viral load (Ct value) in the C+P+ group was 17.2 (13.1–18.8) cycles. There was no difference in estimated viral load or ACE2 receptor and TMPRSS2 expression in the placentas from women affected by COVID-19 with or without transplacental transmission ( Supplemental Table 3 ). The placentas of the 4 C+P+ cases without transplacental transmission were infected by the original SARS-CoV-2 wild type.

General care

Maternal COVID-19 was diagnosed according to the WHO guidance criteria for clinical management, and transplacental transmission was diagnosed according to specific WHO criteria. To diagnose maternal disease, RT-PCR was performed on nasopharyngeal swabs obtained following US Centers for Disease Control and Prevention guidelines. Optimal neonatal care was provided following international guidelines for neonatal resuscitation and respiratory care. ^, Fetal distress was diagnosed and classified according to the American College of Obstetricians and Gynecologists clinical management guidelines. The obstetrical care and follow-up were carried out according to the current national recommendations. All deliveries were performed in full isolation: infants who needed critical care were immediately admitted to negative-pressure isolation NICU rooms and deisolated after at least 1 negative RT-PCR on nasopharyngeal aspirate. Clinical data were collected in real time from maternal files and electronic NICU monitoring system and neonatal files in a dedicated and secure database. Cases of transplacental transmission were described according to CAse REport guidelines.

Sampling techniques

For all cases of transplacental transmission, placental tissue specimens were obtained from the chorionic side and crushed in 400 mL of RNase or DNase-free water. Furthermore, samples were collected from different organs and at different times as follows. Amniotic fluid was sterilely collected during cesarean delivery before spontaneous rupture of membranes. Nonbronchoscopic bronchoalveolar lavage was performed using a standardized procedure previously described and already used to detect SARS-CoV-2 in neonates. This method follows European Respiratory Society guidelines for pediatric and neonatal bronchoalveolar lavage. During the procedure, the patient is not disconnected from the ventilator, and there is usually no major desaturation or bradycardia. Gastric aspirate fluid was sterilely collected immediately after birth using a soft, end-hole suction catheter inserted through the mouth and applying −50 mm Hg pressure. Nasopharyngeal aspirate samples were collected, after having cleaned the neonate with sterile towels to remove vernix caseosa and blood from the skin, as follows: sterile saline solution (0.9% NaCl, 0.5 mL, room temperature) was injected, and a 4F suction catheter was inserted into 1 nostril to a depth equivalent to the distance between the infant’s ear and nostril; vacuum suction (−50 mm Hg) was applied while gently pulling the catheter. This technique was preferred over nasopharyngeal swabs given the small diameter of nostrils in preterm neonates and because it is more accurate than simple swabs in detecting respiratory viruses in infants. Urine was collected with a sterile plastic bag, applied after having washed the genital area with soap and sterile water. Stools were directly collected from diapers using a dedicated sterile spoon and container. Neonatal samples were never blood stained and were all kept at +4°C and tested within 24 hours.

Virology

Viral RNA was extracted from 200 μL of each sample with the NucliSENS easyMag (bioMérieux, Craponne, France) and eluted in 100 μL. The RealStar SARS-CoV-2 RT-PCR Kit 1·0 (limit of detection = 1200 cp/mL [12 cp/rxn]; altona Diagnostics GmbH, Hamburg, Germany) targeting the E (specific for lineage B—betacoronavirus) and the S gene (specific for SARS-CoV-2) were used following the manufacturer’s recommendations. The assay includes a heterologous amplification system (internal positive control) to identify any possible inhibition and reagent validity (reproducibility and inter-assay agreement were 100% for both negative and positive tests). Cycling was performed at 55°C for 20 minutes for reverse transcription, followed by 95°C for 2 minutes and 45 cycles of 95°C for 15 minutes, 55°C for 45 minutes, 72°C for 15 minutes with the ViiA7 device (Applied Biosystems, Thermo Fisher, Waltham, MA). Cycle (Ct) values were interpreted following the European Centre for Disease Control (ECDC) guidelines, and a cutoff value of 35 was used as a positive threshold. Ct was normalized for the number of analyzed cells to estimate the viral load. RT-PCR was performed in duplicate by laboratory investigators (C.V.F. and J.M.J.) blinded to the clinical data. Sanger sequencing of the S gene was performed on all samples and analyzed on SeqScape 4 software (Applied Biosystems, Thermo Fisher, Waltham, MA), following the manufacturer’s recommendations. The software was aligned on reference sequence 20A.EU2 (69–70 and 144 deletions, 501, 484, and 452 mutations on the S gene were particularly investigated), and the analysis followed ECDC guidelines for SARS-CoV-2 sequencing.

Design and patients

This was a translational and prospective cohort study: all pregnant women hospitalized during the study were considered eligible. Patients were recruited if they fulfilled the following criteria: (1) diagnosis of COVID-19 during the third trimester of pregnancy, with negative RT-PCR on placental tissue (C+P− group), (2) diagnosis of COVID-19 during the third trimester of pregnancy, with positive RT-PCR on placental tissue (the 6 aforementioned cases of transplacental transmission were included in this group; C+P+ group), and (3) absence of any SARS-CoV-2 infection (control group). Women for this latter group were selected if they fulfilled all the following conditions: (1) negative history of contact with SARS-CoV-2–infected patients; (2) negative serology during hospital admission, analyzed as previously described ; and (3) negative RT-PCR on nasopharyngeal swab performed as described above. For each recruited patient, RT-PCR was performed on placental specimens (as described above), and specimens were kept at −80°C for further analysis. Clinical data were extracted in real time from electronic patient files and stored in a dedicated database. Pregnancy and neonatal care followed current national and international recommendations for best practice. ^{, ,} Intrauterine growth retardation was diagnosed according to French curves. Fetal distress was diagnosed according to guidelines, as described above. The study received appropriate ethical approval (see below); all relevant regulations were respected, and the Strengthening the Reporting of Observational Studies in Epidemiology guidelines for cohort studies were followed for manuscript preparation.

Molecular biology

Placental tissue was sampled as described above and used to measure the expression of SARS-CoV-2 receptors: ACE2 (Abcam; Discovery Drive, Cambridge, United Kingdom) and TMPRSS2 (Aviva Systems Biology, San Diego, CA) were assayed using commercial highly specific enzyme-linked immunosorbent assay kits validated for human tissue extracts. ^, Their inter-assay and intra-assay coefficients of variation were <12% and <10%, respectively. ACE2 and TMPRSS2 concentrations were corrected for the total protein content in each sample, measured by a colorimetric kit detecting copper reduction by proteins using bicinchoninic acid (Thermo Fisher, Waltham, MA). All assays were performed in duplicate by laboratory investigators (C.V.F. and J.M.J.) blinded to clinical data.

Virology

Placental samples from all study groups were subjected to RT-PCR and viral gene sequencing as described above, by laboratory investigators (C.V.F. and J.M.J.) blinded to clinical data.

Gross examination, histology, and immunohistochemistry

Placental sample preparation and gross and microscopic examination were performed according to the Amsterdam consensus statement. In detail, a minimum of 4 blocks for each placenta were examined (1 block containing a roll of the extraplacental membranes and 2 sections of the umbilical cord and 3 blocks containing full thickness of normal-appearing placenta parenchyma) together with 1 block of each type of lesion. There was no selection of microscopic fields per each biopsy: all slides were fully digitalized and entirely examined as follows.

The placentas were fixed in 4% buffered formalin, and samples were paraffin embedded. Hematoxylin-eosin-saffron stain was used on 3- to 5-μm thick sections. Immunohistochemistry with peroxidase detection and hematoxylin counterstain was performed in a Leica Bond-III automated immune-stainer using the Bond Polymer Refine Detection DS9800 kit (Leica Biosystems, Nanterre, France) after heat pretreatment at pH 6 or 9 depending on the antibodies tested. The following antibodies were used: CD20 (Dako L26, 1:400), CD3 (F7.2.38, 1:50), CD8 (Dako C8-144B, 1:50), CD34 (Dako, QBEnd-10, 1:200), CD68 (Dako PG-M1, 1:200; all from Agilent, Santa Clara, CA), and CD163 (Leica 10D6, 1:200) and CD4 (Leica A4B12; prediluted by the manufacturer, both from Leica Biosystems, Nanterre, France). C4d (A24-T, 1:100; DB Biotech, Kosice, Slovakia), ACE2 (SN0754, 1:300; Genetex International, Hsinchu City, Taiwan), SARS-CoV-2 nucleoprotein (polyclonal, 1:12000; ABclonal Technology, Woburn, MA), and cytomegalovirus CCH2 + DDG9 (Dako M0854, monoclonal, 1:1; Agilent, Santa Clara, CA) were also used. All assays were analyzed in duplicate on digitalized slides (3DHistec Pannoramic 1000 slide scanner, Plan-Apochromat 20× objective and camera adapter 1.6×) with a CaloPix viewer (TRIBVN Healthcare, France) by 2 expert pathologists (S.P. and A.L.B.) blinded to clinical and virology data.

Pathology

The findings of the examination of placentas from control pregnancies were completely normal. Placentas from women diagnosed with COVID-19 (C+P− and C+P+ groups; N=20) showed at least 1 macroscopic or microscopic abnormality that was not specific to any patient (ie, they were not preferentially observed in either the C+P− or the C+P+ group, irrespective of the occurrence of transplacental transmission). These abnormalities were fetal vascular malperfusion (20%), subchorionic thrombosis (35%), intervillous thrombosis (5%), small infarcts (30%), marginal hematoma (20%), and acute chorioamnionitis (45%) with a fetal inflammatory response in one-third of cases. In addition, lymphocytic deciduitis was observed in 20% of the cases, sometimes with several plasma cells.

Conversely, 3 distinct pathologic phenotypes were recognized. The first phenotype was observed in C+P+ cases with transplacental transmission (n=6) and consisted of (1) massive fibrin perivillous deposition >50% of the tissue, (2) diffuse chronic intervillositis with fibrillar and/or dense fibrin deposition with villositis necrosis, and (3) diffuse SARS-CoV-2 nucleoprotein–positive immunohistochemistry in intervillositis lesions. The second phenotype was observed in C+P+ cases without transplacental transmission (n=4) and composed of (1) fibrin perivillous deposits <40% of the placenta with diffuse chronic intervillositis, massive fibrin perivillous deposits >50% of the placenta without chronic intervillositis, or a normal fibrin perivillous amount according to the term with foci of chronic villitis with or without perivillitis and (2) SARS-CoV-2 nucleoprotein–positive immunohistochemistry either as foci at the villi surface surrounded by dense fibrin deposition or as rare positive syncytiotrophoblastic cells. The third phenotype was observed in C+P− cases (n=10) and lacked fibrin perivillous deposition and diffuse chronic intervillositis but presented some foci of chronic perivillitis with fibrin deposits. In this third case, SARS-CoV-2 nucleoprotein immunostaining was either negative or hardly positive on syncytiotrophoblastic cells with fibrin deposits ( Supplemental Figures 3 and 4 ).

SARS-CoV-2 nucleoprotein positivity was detected only in the cytoplasm of syncytiotrophoblastic cells but was absent in villous endothelial cells and Hofbauer cells. In cases with transplacental transmission, the positive signal was very strong in the areas of intervillositis; the intensity of the signal decreased moving away from the inflammatory foci to become negative in the areas devoid of intervillositis ( Supplemental Figure 3 ). Immunostaining for ACE2 receptors confirmed the results described above and showed that these receptors are equally expressed in pregnant women with an infected or uninfected placenta; similarly, C4d deposition was analogous in C+P−, C+P+, and control placentas ( Supplemental Figure 5 ). All placental immunostaining were negative for cytomegalovirus proteins (data not shown).