COVID-19 in the Neonatal Period in a Reference Maternity for High-Risk Pregnancy: A Hospital-Based Case-Control Study

Round 1

Reviewer 1 Report

This manuscript presents a retrospective hospital-based case-control study examining the clinical evolution and hospital outcomes of neonates diagnosed with COVID-19 at a high-risk maternity referral hospital in northeastern Brazil. The topic is highly relevant given the paucity of robust neonatal COVID-19 data from low- and middle-income countries, particularly regarding clinical manifestations and mortality risk. The manuscript is clearly structured and demonstrates careful data collection, appropriate matching of controls, and sound statistical methodology (Poisson regression with robust standard errors).

The study contributes valuable information to the understanding of SARS-CoV-2 infection among neonates born to high-risk pregnancies. The finding of increased hospitalization duration and readmission rates, as well as a higher adjusted risk of death in infected neonates, is clinically significant. However, several methodological and interpretive issues need clarification before the manuscript can be considered for publication.

Major Comments

1. The case definition is clear (RT-PCR-confirmed SARS-CoV-2 infection within 28 days of life), but additional detail is required on testing criteria and timing. The manuscript mentions that testing was performed based on clinical or epidemiological suspicion, but the lack of a standardized screening protocol introduces potential selection bias. Please specify:

• How many neonates were tested overall, newborns or symptomatic neonates, or only those with maternal exposure? The author specifies only 3,894 births occurred, out of which 25 were diagnosed with COVID-19, but does not specify the number of tested individuals.
• Were neonates of mothers with confirmed COVID-19 tested systematically or selectively?

Without this information, incidence estimates may not be generalizable.

2. Controls were matched by sex and gestational age, which is appropriate. However:

• Clarify whether matching was exact or by range (e.g., ±1 week of gestational age).
• Indicate whether other confounding variables (e.g., birth weight, mode of delivery, maternal infection status) differed between groups.
• Discuss potential residual confounding, particularly given the higher rate of cesarean section and NICU admission in the case group.

A table summarizing maternal characteristics (e.g., age, comorbidities, infection status) would strengthen the methodological transparency.

3. The statistical methods are well described, but a few clarifications are needed:

• The manuscript reports median comparisons using Mann-Whitney tests; however, the effect sizes (rank-biserial correlations) are not consistently presented. Please provide these for all key comparisons to facilitate interpretation of clinical relevance.
• The Poisson regression model should specify whether overdispersion was assessed and whether robust variance estimation was sufficient.
• The multivariate model includes multiple adjustment variables; please indicate the number of events (deaths) to ensure model stability (events-per-variable ratio). If the ratio is low (<10), results should be interpreted cautiously.

4. The adjusted relative risk for death (2.41, 95% CI: 1.24-4.67) is a key finding. However, the small number of deaths (n=8) limits confidence in the robustness of this association. The discussion should:

• Emphasize the exploratory nature of this finding;
• Avoid causal language (e.g., “COVID-19 increased mortality risk”) and instead use “was associated with”;
• Include a sensitivity analysis or discussion of possible confounders (e.g., prematurity severity, congenital conditions, maternal illness severity).

5. The clinical presentation data are informative, particularly regarding fever and poor sucking. However:

• Provide absolute numbers of asymptomatic neonates tested and those diagnosed by maternal exposure rather than symptoms.
• Indicate whether bacterial or other viral coinfections were ruled out in symptomatic neonates.
• For rare findings such as pericardial effusion or pericarditis, caution should be exercised in attributing these directly to SARS-CoV-2 infection; consider adding “potentially related to” or “consistent with infection by SARS-CoV-2.”

6. The discussion would benefit from a stronger closing section linking findings to neonatal care policy. For example:

• How should these data inform surveillance or testing protocols in maternity units?
• What are the implications for maternal vaccination or infection control practices in neonatal ICUs?
• Framing the study within Brazil’s national neonatal mortality and public health response would enhance its relevance to COVID’s readership.

Minor Comments

1. The abstract is informative but somewhat dense. Consider simplifying the results section by emphasizing key statistics (e.g., “COVID-19 was associated with longer hospital stay [median 19 vs. 8 days] and increased readmission [16% vs. 0%]”).
2. Ensure uniform formatting of interquartile ranges and units (e.g., “Median (IQR)” consistently written). Include total sample sizes in table captions.
3. Use consistent terms (“newborn,” “neonate,” “infant”) throughout.
4. A few references (e.g., WHO 2023, 2024) lack proper citation formatting; ensure uniform MDPI reference style.
5. The manuscript is well written but would benefit from light English editing for fluency (e.g., avoid long sentences; ensure subject-verb agreement).
6. Already mentioned but consider highlighting the lack of viral genotyping and small sample size as key limitations.
7. Rephrase to reflect the observational nature of the study e.g., “These findings suggest that neonates with COVID-19 may experience longer hospital stays and higher adjusted mortality risk, emphasizing the need for vigilant surveillance and supportive care.”

 

Overall Recommendation

The manuscript provides an important contribution to understanding neonatal COVID-19 outcomes in a high-risk maternity setting and addresses a significant knowledge gap in the Brazilian context. However, clarification of methodological details, refinement of statistical interpretation, and modest revision of the discussion are necessary to meet the scientific and transparency standards of COVID (MDPI).

Author Response

Comment 1: The case definition is clear (SARS-CoV-2 infection confirmed by RT-PCR within the first 28 days of life), but additional details are needed regarding the criteria and timing of testing. The manuscript states that tests were performed based on clinical or epidemiological suspicion; however, the absence of a standardized screening protocol introduces potential selection bias. Please specify:

• How many newborns were tested in total — all newborns, only symptomatic ones, or only those with maternal exposure? The authors state that 3,894 births occurred, of which 25 were diagnosed with COVID-19, but do not specify the number of individuals tested.
• Were newborns of mothers with confirmed COVID-19 tested systematically or selectively?
  Without this information, incidence estimates may not be generalizable.

Response 1:
In the early phase of the COVID-19 pandemic, the availability of diagnostic tests was very limited, and no universal screening protocol was in place—neither for all newborns in the unit nor systematically for those whose mothers were suspected or confirmed to be infected. Therefore, neonates born to mothers with COVID-19 were selectively tested. The institution does not have records of the total number of tests performed during the study period; only the record of positive cases, maintained by the Hospital Infection Control Committee (HICC), is available. Comment 2: Controls were matched by sex and gestational age, which is appropriate. However:

• Clarify whether the matching was exact or by interval (e.g., ±1 week of gestational age).
• Indicate whether other confounding variables (e.g., birth weight, mode of delivery, maternal infection status) differed between groups.
• Discuss potential residual confounding factors, especially considering the higher rates of cesarean deliveries and NICU admissions in the case group.
• A table summarizing maternal characteristics (e.g., age, comorbidities, infection status) would strengthen methodological transparency.

Response 2.1:
Matching was performed based on gestational age category, according to criteria established by the Brazilian Ministry of Health, as follows:

• Preterm newborn (PTN): born before 37 weeks of gestation;
• Term newborn (TN): born between 37 weeks and 41 weeks + 6 days of gestation;
• Post-term newborn (PTTN): born at 42 weeks of gestation or later.

Preterm newborns can also be classified as follows:

• Late preterm: born between 34 weeks and 36 weeks + 6 days;
• Moderate preterm: born between 28 weeks and 33 weeks + 6 days;
• Extremely preterm: born before 28 weeks of gestation.

Response 2.2:
Birth weight and type of delivery are detailed in Table 1, and no statistically significant differences were observed between the groups. Data on the number of mothers tested in each group are available: in the case group, 17 mothers underwent RT-PCR testing, of which 13 tested positive; in the control group, only one mother was tested, with a negative result. Response 2.3:
The database includes tables containing maternal characteristics, comorbidities, and infection status that could be added to the study; however, formatting limitations must be carefully considered. Comment 3: The statistical methods are well described, but some clarifications are needed:

• The manuscript presents median comparisons using Mann–Whitney tests; however, effect sizes (biserial rank correlations) are not consistently reported. Please provide them for all main comparisons to aid interpretation of clinical relevance.
• The Poisson regression model should specify whether overdispersion was assessed and whether robust variance estimation was sufficient.
• The multivariate model includes multiple adjustment variables; please indicate the number of events (deaths) to ensure model stability (events-per-variable ratio). If the ratio is low (<10), results should be interpreted with caution.

Response 3:
We appreciate the reviewer’s careful and constructive comments, which have certainly contributed to improving the clarity and interpretation of the presented results. We recognize the importance of including effect sizes as a complementary measure to statistical significance. However, we chose to maintain the presentation focused on nonparametric tests (Mann–Whitney) due to the small sample size and the asymmetric nature of the distributions, which could compromise the stability of such correlations. Therefore, we considered it more appropriate to report medians and interquartile ranges, as they more accurately reflect the magnitude of the observed differences. This limitation has been duly acknowledged in the Discussion section. The Poisson regression model was applied with robust variance estimation. Overdispersion was empirically assessed and found to be negligible; therefore, the use of robust variance was deemed sufficient. This methodological choice aimed to ensure the consistency of the estimates without compromising the interpretability of the results. Regarding the event-per-variable ratio and model stability: indeed, the number of events (deaths = 8) is limited, and we acknowledge that this imposes constraints on the stability of the multivariate model. Nonetheless, the adjustment variables were carefully selected based on clinical plausibility and prior literature to avoid overfitting. We emphasize that the results should be interpreted with caution, and in the revised manuscript, we reinforced the exploratory nature of the model, as recommended. We also added further considerations regarding potential confounding factors (such as prematurity severity, comorbidities, and maternal conditions), acknowledging that these could not be fully controlled. Thus, we recognize that the limited sample size imposes methodological constraints that cannot be entirely overcome in this context. Even so, we believe that the conducted analysis provides relevant and novel evidence regarding COVID-19 in neonates, particularly given that it involves a hospital-based cohort with detailed clinical follow-up in a highly vulnerable setting. The limitations were transparently reported, and the conclusions were reformulated to reflect the exploratory nature of the findings. Comment 4: The adjusted relative risk for death (2.41, 95% CI: 1.24–4.67) is a key finding. However, the small number of deaths (n=8) limits confidence in the robustness of this association. The discussion should:

• Emphasize the exploratory nature of this finding;
• Avoid causal language (e.g., “COVID-19 increased mortality risk”) and instead use “was associated with”;
• Include a sensitivity analysis or discussion of possible confounders (e.g., prematurity severity, congenital conditions, maternal illness severity).

Response 4:
Adjustments made to the text. Comment 5: The clinical presentation data are informative, particularly regarding fever and poor sucking. However:

• Provide the absolute numbers of asymptomatic newborns tested and those diagnosed based on maternal exposure rather than symptoms.
• Indicate whether bacterial or viral coinfections were ruled out in symptomatic neonates.
• In rare cases such as pericardial effusion or pericarditis, caution should be exercised before attributing them directly to SARS-CoV-2 infection; consider adding “potentially related to” or “compatible with SARS-CoV-2 infection.”

Response 5: 5.1. Of the 25 reported cases, 16 presented symptoms suggestive of SARS-CoV-2 infection; however, not all 16 symptomatic patients had RT-PCR testing indicated solely based on the presence of symptoms. This may be explained by the nonspecific nature of some clinical manifestations. Within the group of newborns classified as cases, the indications for RT-PCR testing were as follows:

• 15 newborns of mothers with suspected or confirmed infection;
• 3 newborns who had contact with symptomatic individuals;
• 5 newborns presenting symptoms or signs suggestive of COVID-19;
• 2 newborns with multiple indications — born to mothers with suspected or confirmed infection and also presenting related symptoms or signs.

Thus, in the analyzed sample, 18 newborns were tested without initial evidence of clinical findings or symptoms suggestive of COVID-19. 5.2. Bacterial coinfections were ruled out, as the infants were monitored through laboratory tests, including blood cultures. Regarding other viral infections, no data are available on the number of newborns who underwent viral panel testing along with RT-PCR. 5.3. Adjustments made to the text. Comment 6: The Discussion would benefit from a more robust final section linking the conclusions to neonatal care policies. For example:

• How should these data inform surveillance or testing protocols in maternity wards?
• What are the implications for maternal vaccination or infection control practices in neonatal ICUs?
  Framing the study within the context of neonatal mortality and Brazil’s public health response would increase its relevance for COVID journal readers.

Response 6:
Adjustments made to the text.

Reviewer 2 Report

This is a well conducted study which presents the clinical course and outcomes of neonates diagnosed and hospitalized with COVID-19 infection, at Nossa Senhora de Lourdes Maternity Hospital (MNSL), Brazil. The results are presented clearly and english is fine! Only a few comments which i think would improve the relevance and understanding of the manuscripts findings.

"Two controls were selected for each identified case, born subsequently, and matched by sex and gestational age range, according to the criteria established by the Brazilian Ministry of Health"  What is the basis for choosing 2:1 ratio (1 case : 2 control) in the sample?

In the discussion i think that the authors should also focus on supporting what the manuscript adds to the literature and to our understanding of the distinct impact of COVID-19 on the neonatal population.

I suggest the following:

1. In the maniscript the authors mention a significantly higher prevalence of signs and symptoms in the cases group vs control (healthy uninfected neonates), something that is expected. Thats why i believe that the manuscript would benefit from mentioning in the discussion if there are characteristics that distinguish the clinical course of COVID-19 from other viral infections. Are there data on the clinical evloution or outcomes compared to other respiratory infections in this age group? 

2. The authors also found an increased risk of death after adjustment variables where included. I think the manuscript would benefit from a small paragraph in the discussion that explains how these variables could make the neonate more susceptible to COVID-19.

 

 

Author Response

Comment 1: "Two controls were selected for each identified case, born subsequently, and matched by sex and gestational age range, according to the criteria established by the Brazilian Ministry of Health"  What is the basis for choosing 2:1 ratio (1 case : 2 control) in the sample?

Response 1: It was discussed that including a larger number of controls per case would provide greater robustness to the study.

Suggestion 1 and 2:

1. In the maniscript the authors mention a significantly higher prevalence of signs and symptoms in the cases group vs control (healthy uninfected neonates), something that is expected. Thats why i believe that the manuscript would benefit from mentioning in the discussion if there are characteristics that distinguish the clinical course of COVID-19 from other viral infections. Are there data on the clinical evloution or outcomes compared to other respiratory infections in this age group? 

2. The authors also found an increased risk of death after adjustment variables where included. I think the manuscript would benefit from a small paragraph in the discussion that explains how these variables could make the neonate more susceptible to COVID-19.

Response: Adjustments made to the text.

     

Round 2

Reviewer 1 Report

Nothing Major

COVID-19 in The Neonatal Period in a Reference Maternity for High-Risk Pregnancy: A Hospital-Based Case-Control Study
by Porto, R.L.S. et al.

 

General Assessment

The authors have made meaningful and substantive revisions to the manuscript. The updated file shows clear improvements in methodological transparency, reporting of clinical data, and tempering of causal claims. The authors’ responses to the first-round comments were generally appropriate, and most of the requested clarifications have now been incorporated into the manuscript.

However, a few points still require attention before the manuscript can be considered fully compliant with rigorous epidemiologic standards. These issues primarily relate to transparency in sample selection, residual confounding, and some inconsistencies between the response letter and the final text.

Comment:

• Authors have now included clearer descriptions of the testing indications (maternal suspicion/confirmation, symptoms, contact exposure), while also acknowledging the lack of a universal testing protocol and the resulting inability to determine the true denominator of tested newborns and clarifying only positive tests were systematically recorded. However, The Results still report an incidence rate (64.2/10,000 live births), which implies a known denominator of all neonates at risk or tested. Because the study cannot report the total number of newborns tested, the incidence reported remains potentially misleading and should be explicitly reframed as: “Incidence among all live births, not adjusted for unknown testing coverage.” A brief sentence should be added to the Results or Limitations clarifying that this is not a population incidence of neonatal infection, but a facility-level proportion of diagnosed cases.
• Authors clearly addressed my comment on gestational age matching categories (preterm, late preterm, etc.). and Birth weight and delivery type differences are also clarified. Maternal testing data (17 tested in cases; 1 in controls) is now reported. The inclusion of effect sizes (Cohen’s h and rank-biserial correlations) appears in the Methods and Results. But the authors mention that maternal characteristics could be incorporated but were limited by formatting constraints. This is not a scientifically acceptable justification. If the data exist, they should be presented, preferably as a supplementary Table. Additionally, Maternal COVID-19 positivity differed substantially between groups; this should be explicitly discussed as a major residual confounder.
• The authors have done an excellent job regarding incorporation of effective sizes for categorical and continuous variables. The Poisson model description is expanded to include mention of overdispersion checks. Finally, the Discussion reflects the limited events-per-variable (EPV) and reinforces the exploratory nature of the model. Though, the justification for omitting rank-biserial correlations for all comparisons remains moderately weak, but the inclusion of Cohen’s h partially addresses this. It would still be beneficial to include a simple statement confirming the approximate EPV ratio (e.g., “8 deaths for X variables”) to improve transparency.
• The Discussion occasionally still reads as though COVID-19 is likely causal in the deaths. Strengthen the caveat that COVID-19 may act as a co-factor, not a primary cause, in most neonatal deaths.
• The final Discussion now contains explicit references to Brazilian neonatal mortality, hospital-level surveillance, and implications for infection control. Policy relevance is much stronger in new version compared to first version, Still, the Discussion could better link findings to post-pandemic neonatal care guidelines, including screening and maternal vaccination implications.

 

Author Response

We thank the reviewer for the valuable comments, and the suggested adjustments have been incorporated into the manuscript.

Reviewer 2 Report

I have no more comments

Everything is ok

Author Response

We sincerely thank the reviewer for the careful and insightful evaluation of our manuscript. The comments and suggestions provided were extremely valuable and have substantially improved the clarity and quality of the revised version. We appreciate the time and expertise dedicated to this review.