Study design and participants

All pregnant women diagnosed with COVID-19 who gave birth between January 30 to March 10, 2020 in 2 designated general hospitals in Wuhan, China (Renmin Hospital, Wuhan University, and Central Hospital of Wuhan, Tongji Medical College, Huazhong University of Science and Technology) were included in this retrospective study. Clinical characteristics, examination results, and treatment course were extracted from their medical records.

All procedures performed in this study involving human participants were in accordance with the ethical standards of the Medical Ethical Committee of Wuhan Children’s Hospital (Wuhan Maternal and Child Healthcare Hospital), Tongji Medical College, Huazhong University of Science and Technology (WHCH2020014), Central Hospital of Wuhan, Tongji Medical College, Huazhong University of Science and Technology (2020–28), and Renmin Hospital, Wuhan University (WDRY2020-K097). For this retrospective study in clinical practice, the requirement for written informed consent was waived by the ethics committee.

Maternal data

We followed the diagnostic criteria for COVID-19 in pregnant women according to the New Coronavirus Pneumonia Prevention and Control Program (fifth, sixth, and seventh edition) [15–17] published by the National Health Commission of China (specific diagnostic criteria were shown in S1 Text). In women where a PCR-kit for SARS-CoV-2 RNA was not available, the diagnosis was made with chest CT scan.

For mothers, we collected sociodemographic data (maternal age, body mass index [BMI], educational attainment [>high school or not], and occupation [employed specifically at a hospital, employed but not at a hospital, or unemployed]), maternal parity (primiparous or multiparous), medical history records (fever, cough, shortness of breath, diarrhea, and vomiting), laboratory tests, chest CT images, throat swab tests, use of mechanical ventilation, and intensive care unit [ICU] admission. Laboratory tests included routine antenatal blood tests (white blood cell count [WBC], lymphocyte count [LYM], lymphocyte percentage [LYM%]), levels of C-reactive protein (CRP) in serum, biochemical indicators of hepatic and renal function (aspartate transaminase [AST], alanine aminotransferase [ALT], creatine kinase [CK], lactate dehydrogenase [LDH], total protein [TP], albumin [ALB], uric acid [UA], creatinine [CREA], and urea nitrogen [BUN]), and postnatal blood routine tests and levels of CRP. For all women, we registered pregnancy outcomes, i.e., gestational age at delivery, mode of delivery, and occurrence of gestational hypertensive disorder, gestational diabetes mellitus, premature rupture of membranes, and fetal distress.

Neonatal data

Neonates who need hospital care or who were admitted on parental request for quarantined observation were admitted to a designated children’s hospital (Wuhan Children’s Hospital [Wuhan Maternal and Child Healthcare Hospital], Tongji Medical College, Huazhong University of Science and Technology). When parents did not opt for quarantined observation, neonates were discharged after birth. For these neonates, the outcome was obtained through parental telephone interview.

Neonatal data collected included sex, birthweight, 1- and 5-minute Apgar scores, congenital anomalies, fever, respiratory distress and neonatal ICU (NICU) admission. Neonates admitted at NICU care had additional data collected including throat swabs, anal swabs, as well as chest X-rays or CT images. Laboratory data collected included routine blood tests (WBC, LYM, LYM%, neutrophil count [NEU], neutrophil percentage [NEU%], platelet count [PLT]), levels of CRP, procalcitonin (PCT) in serum, and biochemical indicators of hepatic and renal function (AST, ALT, CK, LDH, TP, ALB, UA, CREA, and BUN). In addition, according to the New Coronavirus Pneumonia Prevention and Control Program (seventh edition) [17], SARS-CoV-2 specific immunoglobulin M (IgM) and immunoglobulin G (IgG) was examined in 4 neonates’ serum. The diagnosis of pneumonia in neonates was based on risk factors, clinical manifestations, radiological findings, and laboratory tests [18]. The diagnosis of COVID-19 in neonates was according to an experts′ consensus statement [19] (specific diagnostic criteria are shown in S1 Text).

SARS-CoV-2 testing

Maternal and neonatal throat swabs and neonatal anal swabs were collected and tested for SARS-CoV-2 RNA using quantitative real time-PCR (qRT-PCR) technology and hydrolysis probe technology (the New Coronavirus 2019 Nucleic Acid Detection Kit [Dual Fluorescence PCR] provided by Shuo Shi Biotechnology Co., Ltd, Jiangsu, China). SARS-CoV-2 specific IgM and IgG were detected using New Coronavirus 2019-nCoV IgG Antibody Detection Kit (Chemiluminescence) (Yahuilong Biological Technology Co., Ltd. Shenzhen, China) [20–22]. Sample collection, processing, and laboratory testing followed guidance from the World Health Organization [23].

Statistical analysis

All analyses were performed using SAS 9.3 (SAS Institute, Inc, North Carolina, USA, https://www.sas.com/zh_cn/software/university-edition/download-software.html#windows). Continuous variables were described as mean (standard deviation [SD]) or median (interquartile range [IQR]). Categorical variables were represented as frequencies with proportions. All the research data were available at the ResMen Manager of Chinese Clinical Trial Registry (www.medresman.org), and the registration number is ChiCTR2000031954.

Main findings

In this study, among 29 women with confirmed or clinical diagnosed COVID-19 infection, 5 neonates were diagnosed with a COVID-19 infection (2 confirmed and 3 suspected). None of these 5 neonates had contact with infected patients, except for their mothers during delivery, which suggests vertical and intrapartum transmission.

Results in context

Several case series have been reported focusing on the impact of COVID-19 on pregnant women and neonates [6–12]. Chen and colleagues [6] first evaluated the clinical characteristics of COVID-19 infection in 9 mother–neonate pairs. Amniotic fluid, cord blood, neonatal throat swab, and breastmilk samples were tested from 6 patients for SARS-CoV-2, and all were negative. In these case series or case reports, the majority COVID-19-infected pregnant women had mild symptoms and were delivered by CS. Further, all women had lung radiological changes, which is consistent with our results. Although in the study by Zhu and colleagues, 10 neonates experienced several clinical problems including premature labor, shortness of breath, thrombocytopenia, and even death [9]; in other reports, positive neonatal outcomes were minimal. Only 2 cases of confirmed neonatal COVID-19 infection have been reported [24]. One neonate was diagnosed at 17 days postdelivery after close contact with a COVID-19-infected mother and babysitter [25]. The other neonate was found to be infected 36 hours after a CS delivery after negative cord blood, amniotic fluid, and placenta SARS-CoV-2 RNA results [26].

In this present study, all neonates with COVID-19 infection were born via CS where both general hospitals had only 1 negative-pressure operating room. When the neonates were born through CS, they stayed unprotected in the operating room for about 30 minutes for observation before they were transferred to isolated neonatal wards. Although a guideline released by the Hubei government required immediate isolation of neonates from infected mothers [27], the 30 minutes’ observation had not been noticed by obstetricians. The contaminated air in the operation room might be the source of the viral transmission, although other transmission routes cannot be excluded.

Strengths and limitations

This is the first study to our knowledge to integrate the complete data of mothers and neonates to explore the association between maternal COVID-19 infection and neonatal COVID-19 infection or other health conditions. All women with confirmed or clinical diagnosed COVID-19 infections who gave birth at 2 of the 5 designated general hospitals between January 30 to March 10, 2020 during the most severe time of the epidemic in China were included in this study. The detailed clinical course of each hospitalized neonates was recorded and the condition of the nonhospitalized neonates was monitored by telephone interview, thus providing reliable follow-up data, which have eliminated the recall bias often found in retrospective studies. However, in the early stage of the outbreak, there was a shortage of COVID-19 test kits. As such, most neonates were only tested twice with throat and anal swabs. Only 4 neonates were tested for SARS-CoV-2 specific IgM and IgG. Ai and colleagues [28] reported chest CT had higher sensitivity in diagnosing a COVID-19 infection in comparison to qRT-PCR tests from swab samples in China. Although more than half of the pregnant women in this study were clinically diagnosed based on chest CT images due to the lack of PCR kits, we believe the diagnosis was accurate. Another limitation is the lack of tracheal aspirate to identify specific pathogens from the first 12 neonates with radiological features for pneumonia. This led to the uncertain etiology of the identified pneumonia. Because we found only 5 neonates with (suspected) COVID-19 infection, comparison between groups had limited statistical power.

Research implications

Our findings suggest maternal CRP levels may be associated with neonatal lung radiological changes, as antenatal CRP levels were higher in mothers of neonates who had abnormal radiological findings (with COVID-19 or without). CRP is the most commonly researched acute-phase reactant in infection and noninfectious inflammation. There is some evidence to suggest antenatal CRP levels are associated with pregnancy-related complications [29], mental disorders in offspring [30], infant birthweight [31], and neonatal CRP levels [32]. Although CRP cannot be transmitted to the fetus through the maternal–fetal interface, maternal inflammation can affect the fetal immune state including immune cells, cytokines, and chemokines [33].

In the 5 COVID-19-infected neonates in our study (Table 4), we found their clinical manifestations were very different from those of infected adults and even children [19]. Neonates had almost no symptoms, and it was difficult to judge the changes in white blood cells and lymphocytes because of the differing number of days since birth. In terms of the SARS-CoV-2 RNA tests in throat and anal swab samples, there may have been sampling errors and false negative or false positive results. As for the detection of the SARS-CoV-2 specific IgM and IgG, only 4 cases were tested in this study due to late arrival of the testing platform, which may become a powerful method for identifying the infections in the future. When comparing all the various diagnostic examinations, the chest CT images appear to be the most important for a diagnosis and prognosis of COVID-19 among neonates. The typical radiographic findings of COVID-19 pneumonia in those 5 infants were predominately manifested as focal or diffuse GGOs located in the peripheral regions, which were distinguishable from bacterial infection but nonspecific from other virus pneumonia, such as influenza A. After treatment, all neonates showed improvement on chest X-ray or CT images. However, in 1 neonate (Patient 18), we found the densities of previous existing lung lesions increased with interlobular septal thickening, suggested that interstitial fibrosis had developed and may influence pulmonary function in the future. As previously reported [34], CT scan is more sensitive than X-ray film in detecting pulmonary lesions which are usually overlapped by the shadow of the heart or hidden behind posterior region of the diaphragms. Considering the safety of ionizing and the need of avoid wasting a medical resource, it is urgent to develop an effective standard to screen infants and children for COVID-19.

Notably, although 12 of the 13 other hospitalized neonates presented radiological features for pneumonia through X-ray or CT, only 1 neonate had an occasional cough, and none of them had any of the usual symptoms associated with a SARS-CoV-2 infection. It seems that neonates born to mothers with COVID-19 had a significantly increased risk for pneumonia-like damage within 6 days of birth. Generally, the incidence rate of neonatal pneumonia is <1% among full-term infants and approximately 10% in preterm infants [35,36]. However, the neonates in this study had no specific symptoms of pneumonia, so no microbiological test had been implemented except for SARS-CoV-2. The increased radiological features for pneumonia may be due to false negative of SARS-CoV-2 PCR test, which has a 30% false negative rate [37]. For example, Patient 12 had a typical CT feature for COVID-19 infection, but her throat and anal swabs were all negative. As such, it is possible that some of these 12 neonates had a potential SARS-CoV-2 infection and that early tracheal aspirate should be obtained for specific pathogen identification. Further, inhalation of amniotic fluid and neonatal wet lung syndrome may also cause abnormal radiological findings. Following treatment, all the lesions shown in S1 Fig resolved, but it is unknown if any long-term effects will develop and warrants investigation.

For SARS-CoV-2 specific serum IgM and IgG, 2 neonates were positive (Patient 10 and 18) raising the concern of intrauterine transmission. Currently, SARS-CoV-2 is thought to primarily exist in the respiratory tract and to a lesser extent in the digestive system. Like other respiratory viruses, the positive rate of nucleic acid testing in blood is not high. The vertical transmission from mother-to-fetus via placenta usually requires a high replication level of virus in the blood. This may be the reason why no SARS-CoV-2 was found in the cord blood, amniotic fluid, or placenta as they were in previous studies [6,38]. When the SARS-CoV-2 specific IgM and IgG test in serum was proposed, neonates who had not been discharged were tested. IgM is the earliest synthesized and secreted antibody during human development. The fetus at the late stage of development begins to synthesize IgM, which is too large to be passed through the placenta and into the fetus. The rise of IgM is generally considered to be an indicator of intrauterine or perinatal infection. On the other hand, IgG appears to be passively transported from the mother to the fetus because it can pass through the placenta and can be synthesized by the infant after 3 months of age [39]. It is unfortunate both cases had no retained cord blood, placenta, or amniotic fluid samples so that we could test for the antibodies. Additional research is needed to further clarify vertical transmission.

In addition, 3 neonates developed NEC despite the lack of prematurity. Similar findings were found among the 5 infants born to infected women during the severe acute respiratory syndrome (SARS) outbreak [40]. One of the 5 infants born to SARS-infected women developed neonatal respiratory distress syndrome (NRDS) and jejunal perforation, and another developed NEC with ileal perforation shortly after birth. These 2 cases included extremely premature infants in which one was born at 26 weeks’ gestation, and the other was born at 28 weeks. The occurrence of NEC in this study is concerning and warrants further investigation.