The Longitudinal Physiological Reference Values of Middle Cerebral Artery Blood Flow Velocity in Extremely Preterm Infants: A Study Utilizing Multimodal Cerebral Hemodynamic Monitors

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The manuscript addresses a clinically relevant and timely topic in neonatal hemodynamic monitoring. A major strength of the study is the use of a multimodal assessment of cardiovascular and cerebral physiology in extremely preterm infants, integrating Doppler ultrasonography, near-infrared spectroscopy, electrical cardiometry, targeted echocardiography, and conventional clinical parameters.

This multiparametric approach is valuable because cerebral perfusion in extremely preterm infants cannot be adequately interpreted using a single variable such as blood pressure alone. By combining systemic hemodynamic assessment, cardiac output estimation, cerebral oxygenation monitoring, and Doppler-derived cerebral blood flow parameters, the authors provide a more comprehensive physiological framework for evaluating cerebral circulation in this vulnerable population.

The longitudinal design and repeated measurements across postnatal maturation are additional strengths, as they allow the authors to explore how cerebral Doppler parameters evolve in relation to increasing postmenstrual age and body weight. This represents an important contribution to the understanding of cerebral hemodynamic maturation in extremely preterm infants.

Overall, the study has merit and addresses an important clinical question. However, several methodological and interpretative issues limit the strength of the conclusions and currently preclude acceptance for publication. In particular, clarification of the study population definition, justification of the physiological inclusion criteria, refinement of the statistical analyses, and a more cautious interpretation of the findings are required. Substantial major revisions are therefore necessary before the manuscript can be further considered for publication.

 

#Comment 1

The criteria used to define hemodynamically significant patent ductus arteriosus (hsPDA) require reconsideration. The manuscript defines hsPDA as "any right-to-left shunt, ductal diameter >2.0 mm, or left atrium-to-aortic root ratio >1.4." However, this definition does not reflect the current understanding of ductal hemodynamic significance in extremely preterm infants.

First, right-to-left ductal shunting is not a criterion of hsPDA but rather suggests elevated pulmonary vascular resistance or pulmonary hypertension during transitional circulation and should not be used as an isolated marker of ductal significance. Furthermore, ductal diameter and LA/Ao ratio alone are insufficient to establish hemodynamic significance, as current recommendations emphasize an integrated echocardiographic assessment that combines multiple indices of pulmonary overcirculation and systemic hypoperfusion.

Recent consensus statements recommend that hsPDA should be defined using a comprehensive echocardiographic evaluation including ductal diameter, shunt direction and velocity, left heart volume loading (LA/Ao ratio and left ventricular output), evidence of systemic steal (e.g., absent or reversed diastolic flow in the descending aorta or major systemic arteries), and the overall clinical condition of the infant rather than isolated measurements.

Because hsPDA constitutes one of the exclusion criteria used to define the "normal CBF" cohort, the use of a simplified and potentially inaccurate definition may have resulted in misclassification of study subjects. The authors should therefore repeat the analysis using a contemporary, guideline-supported definition of hsPDA or, at minimum, perform a sensitivity analysis to evaluate whether the reported reference ranges remain unchanged after applying updated echocardiographic criteria.

 

 

#Comment 2

The use of a fixed cerebral regional oxygen saturation (rScO₂) range of 65–85% as an inclusion criterion for defining "normal cerebral blood flow" requires further justification.

Current evidence indicates that cerebral NIRS values in preterm infants are not constant across the neonatal period but vary according to gestational age, postnatal age, transitional physiology, and the NIRS device and sensor used. Therefore, a single fixed threshold cannot be considered universally applicable across the heterogeneous population included in this study.

Recently published systematic reviews concluded that reference values should preferably be based on gestational age- and postnatal age-specific centile charts, rather than fixed normal ranges. The review identified high-quality centile curves for cerebral oxygenation during both the immediate postnatal transition and the first 72 hours after birth and emphasized that interpretation of NIRS values should account for gestational age, postnatal age, and the specific NIRS technology employed. Furthermore, the authors explicitly recommend the use of centile charts over fixed "normal values" whenever available.

Given that cerebral rScO₂ constitutes one of the principal criteria used to define the study reference population, the application of a uniform threshold (65–85%) across infants with different gestational ages and postnatal ages may have introduced selection bias and misclassification of physiologically normal observations. Infants who are entirely normal according to published gestational age-specific centile charts could have been excluded, whereas others might have been included despite values outside the expected physiological distribution for their developmental stage.

The authors should therefore provide a robust justification for selecting this fixed threshold and discuss its limitations in light of current evidence. If gestational age- and postnatal age-specific NIRS reference curves are available for their cohort, a sensitivity analysis using contemporary centile-based criteria would substantially strengthen the validity of the proposed cerebral blood flow reference ranges.

 

#Comment 3

The statistical methodology requires additional clarification.

Although generalized estimating equations (GEE) are appropriate for repeated measurements, several methodological details remain insufficiently described. The manuscript does not explain the rationale for choosing an independence working correlation matrix, nor does it discuss whether alternative correlation structures were evaluated.

Furthermore, only associations between Doppler parameters and PMA, weight, and blood pressure are presented. Other clinically relevant determinants of cerebral blood flow—including postnatal age, respiratory support, carbon dioxide levels, hemoglobin concentration, and vasoactive treatment—are not considered or discussed. Even if these variables were unavailable because of the study design, their potential confounding effect should be acknowledged.

Finally, the manuscript should clarify how repeated observations from the same infant were balanced across developmental stages and whether missing measurements were handled according to predefined statistical procedures.

 

#Comment 4

An apparent inconsistency exists between the Results section and the statistical results presented in Table 3, which should be clarified before the findings can be adequately interpreted.

The Results state that "On the other hand, EDV, PI and RI showed no significant correlation to PMA and weight." However, Table 3 reports statistically significant P values (P < 0.001) for EDV, PI and RI across weight groups. This discrepancy creates uncertainty regarding the interpretation of the analyses and the actual relationship between Doppler-derived parameters and maturation.

It is unclear whether the authors are referring to two different statistical approaches—for example, an overall comparison among age groups presented in Table 3 versus the longitudinal generalized estimating equation (GEE) analysis—or whether this represents an inconsistency in reporting. If different statistical models were used, this should be explicitly stated in both the Methods and the Results, together with a clear explanation of why apparently significant univariate comparisons became non-significant after longitudinal modeling.

In addition, Table 3 currently reports only P values without effect estimates. Statistical significance alone does not allow readers to appreciate the magnitude or clinical relevance of the observed associations. Reporting regression coefficients (β estimates), confidence intervals, and corresponding P values for the GEE models would considerably improve the transparency and interpretability of the results.

The apparent contradiction between the narrative description and the tabulated results should therefore be resolved, as it currently prevents readers from understanding which conclusions regarding EDV, PI and RI are supported by the statistical analysis.

 

#Comment 5

The main limitation of the study lies not in the quality of the longitudinal dataset but in the interpretation of its findings.

The authors conclude that the study establishes reference values for cerebral blood flow parameters in extremely preterm infants. However, these proposed reference values are derived from only 40 individual infants, albeit with repeated longitudinal measurements. While repeated observations appropriately characterize within-subject physiological maturation, they do not substantially increase the biological variability represented by the study population. Consequently, the cohort may not be sufficiently large or representative to support the establishment of broadly applicable normative reference values.

In contrast, the longitudinal design represents one of the principal strengths of the study. The repeated Doppler assessments provide valuable information regarding the physiological evolution of cerebral blood flow parameters during postnatal maturation and demonstrate how these parameters change in relation to increasing postmenstrual age and body weight.

For this reason, the manuscript would be strengthened by emphasizing the developmental trajectories of cerebral Doppler parameters rather than presenting the findings as definitive reference values. The observed longitudinal changes according to postmenstrual age and body weight constitute an important physiological contribution and are well supported by the study design. Conversely, establishing universally applicable reference ranges would require validation in substantially larger and more heterogeneous populations.

Accordingly, the authors should consider revising the title, objectives, discussion, and conclusions to focus on the maturation-related evolution of cerebral blood flow parameters in physiologically stable extremely preterm infants, rather than suggesting the establishment of definitive normative reference values. Such an interpretation would better reflect both the strengths and the limitations of the available data and would avoid overstating the clinical applicability of the findings.

#Comment 6

The definition of the study population does not include arterial carbon dioxide tension (PaCO₂), despite its well-established role as one of the principal physiological determinants of cerebral blood flow in preterm infants.

PaCO₂ is a potent regulator of cerebrovascular tone. Both hypocapnia and hypercapnia can induce rapid and clinically relevant changes in cerebral blood flow through cerebral vasoconstriction and vasodilation, respectively. In extremely preterm infants, these effects may be even more pronounced because cerebrovascular autoregulation is frequently immature and highly susceptible to changes in arterial carbon dioxide tension. Consequently, cerebral Doppler velocities may vary substantially despite otherwise stable systemic hemodynamic conditions.

This issue is particularly relevant because the aim of the present study is to identify a cohort with "normal cerebral blood flow". It is conceivable that infants fulfilling all predefined hemodynamic, echocardiographic and NIRS criteria nevertheless exhibited significant variations in PaCO₂ capable of influencing cerebral blood flow independently of postmenstrual age or body weight. Therefore, differences in Doppler parameters may partly reflect respiratory physiology rather than physiological maturation alone.

Recent reviews on neonatal cerebral autoregulation consistently recognize PaCO₂ as one of the major determinants of cerebral perfusion and emphasize that fluctuations in carbon dioxide should always be considered when interpreting cerebral hemodynamic measurements. In addition, current recommendations on neonatal hemodynamic assessment acknowledge that cerebral blood flow cannot be interpreted independently of respiratory variables, particularly carbon dioxide, because cerebral vascular reactivity remains largely preserved even when pressure autoregulation is impaired.

The authors should clarify whether arterial or capillary blood gas measurements were available at the time of each Doppler examination and discuss the potential confounding effect of PaCO₂ on the reported findings. If blood gas data were collected, adjustment for PaCO₂ or, alternatively, exclusion of measurements obtained during significant hypo- or hypercapnia would considerably strengthen the physiological definition of the study cohort. If these data were unavailable, this limitation should be explicitly acknowledged, as residual confounding related to carbon dioxide may have influenced the proposed cerebral blood flow reference values.

Comments on the Quality of English Language

The manuscript is generally well written and easy to follow. Nevertheless, the text contains several grammatical and stylistic inaccuracies, as well as occasional non-native English expressions. Although these issues do not prevent comprehension, careful editing by a native English speaker or a professional language editing service is recommended to improve the overall quality and readability of the manuscript prior to publication.

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

The study addresses a genuine gap widely cited neonatal CBF-velocity norms (e.g., Romagnoli 2006) predate current respiratory, CO₂, and saturation practice, and a contemporary, physiologically anchored dataset is useful. The multimodal definition of stability is novel. The introduction is clear and the clinical framing is strong. The points below are intended to be actionable.

Major comments:

1. Text–table contradiction for EDV, PI, RI (must resolve): The Results (l. 173–175) and Abstract state that EDV, PI, and RI showed no significant association with PMA and weight. This matches Table 2 (PMA: EDV p=0.08, PI p=0.41, RI p=0.16) but directly contradicts Table 3 (weight: EDV, PI, and RI all p<0.01). Please recompute, state which analysis is correct, and align the abstract, text, and both tables. This single inconsistency currently undermines a central conclusion (that only velocity indices are maturation-dependent).
2. CBF vs. CBF velocity: The study measures MCA Doppler velocities, not volumetric flow (mL/min or mL/100 g/min). Velocity is a surrogate that depends on vessel diameter and insonation angle. The title ("Reference of Cerebral Blood Flow"), abstract, and much of the text should be reframed in terms of cerebral blood flow velocity (CBFV) to avoid overstating what was measured. 
3. Definition of "mean velocity" (MV): The appendix waveform (Figure A1) reports both TAmax (23.6) and TAmean (14.3); back-calculating the machine PI of 1.14 from PSV 39.2/EDV 12.1 requires MV ~ 23.6, i.e., the time-averaged maximum (TAMAX), not TAMEAN. If MV was in fact TAMAX, please state this explicitly, because the "mean velocity" label is ambiguous and directly affects PI/RI values and comparability with prior series (many of which use TAMEAN). 
4. GEE reporting is incomplete: Please clarify whether models were univariable or multivariable.
5. Reliability data missing: The text asserts "good inter-operator reliability" across three sonographers and serial timepoints, but no k data are provided. 
6. Methods–Figure discrepancies (reconcile). (a) IVH exclusion threshold: Methods state grade ≥2, Figure 2 states ≥3. (b) Anemia criteria: Methods criteria lists "no RBC transfusion in the preceding week," whereas Figure 2 lists "hemoglobin <10 g/dL"- these are different rules; state which was applied. (c) GA cutoff: Figure 2 shows ≤28+6 weeks vs. ≤28 weeks elsewhere. Please make the stability/exclusion definitions identical across Methods, Figure 2, and the abstract.
7. Comparison with prior norms: The rationale rests on Romagnoli-era values being outdated, yet no quantitative comparison to Romagnoli/Pezzati/Forster is presented. A summary table or overlay contrasting your values with prior series would be required

Minor comments

1. Clarify that weight in Table 3 is current weight at measurement (spanning to ≥2501 g), not birth weight
2. State the rationale for using the left MCA only and whether right-sided or bilateral sampling was considered.
3. EC-derived CO has known limitations in the presence of ductal/atrial shunting; given the PDA burden (ligation 27.5%), briefly address CO measurement validity as part of the stability definition.
4. Table 1: two infants developed IVH ≥2 despite the pre-enrollment exclusion - clarify that post-enrollment IVH datasets were excluded from the "normal" set.
5. Discussion clinical-interpretation paragraph overreaches relative to the data (the study did not examine abnormal states or outcomes). Several sentences are also unclear - e.g., "reduced EDV may indicate compromised cerebral edema" and "Abnormal MV should concern both PSV and MV synergic effect" - please rewrite and frame these as hypotheses rather than validated thresholds.

Comments on the Quality of English Language

Recurring typos: "reginal" (line 30), "Institute of of Emergency" (line 12), "Multidiscipline" in the title (consider "multidisciplinary" or "multimodal"), and inconsistent PMA/weight decimal formatting in Tables 2- 3. A professional English-language edit is recommended; the introduction reads well but the discussion needs tightening.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf