Time to Regain Birthweight and Association with Neurodevelopmental Outcomes among Extremely Preterm Newborns

Objective Determine association between time to regain birthweight and 2-year neurodevelopment among extremely preterm (EP) newborns. Study Design: Secondary analysis of the Preterm Erythropoietin Neuroprotection Trial evaluating time to regain birthweight, time from birth to weight nadir, time from nadir to regain birthweight, and cumulative weight loss with 2-year corrected Bayley Scales of Infant and Toddler Development 3rd edition. Results Among n = 654 EP neonates, those with shorter nadir-to-regain had lower cognitive scores (2–4 days versus ≥ 8 days: −3.5, [CI −7.0, 0.0]; ≤1 day versus ≥ 8 days: −5.0, [CI −10.2, 0.0]) in fully adjusted stepwise forward regression modeling. Increasingly cumulative weight loss was associated with lower cognitive scores (−50 to <−23 percent-days: −4.0, [95% CI −7.6, −0.4]) and language scores (≤−50 percent-days: −5.7, [CI −9.8, −1.6]; −50 to <−23 percent-days: −6.1, [CI −10.2, −2.0]). Conclusion Faster nadir-to-regain and prolonged, severe weight loss are associated with adverse 2-year neurodevelopmental outcomes. Trial registration PENUT Trial Registration: NCT01378273. https://clinicaltrials.gov/ct2/show/NCT01378273

][9][10][11][12][13][14][15][16] Degree of weight loss is associated with time-to-regain, and those with minimal weight loss often regain birthweight earlier.Critically ill premature newborns are commonly provided high-volume uid resuscitation that limit or prevent weight loss after birth and shorten time-to-regain.7][18] Thus, it is possible that a rapid time-to-regain is associated with adverse outcomes.These ndings con ict with recently published recommendations de ning time-to-regain more than 14 days as being associated with malnutrition, implying a shortened time-to-regain is optimal. 19ese con icting recommendations pose unanswered questions.Speci cally, what is the optimal time-toregain among EP newborns to optimize long-term neurodevelopmental outcomes?Time-to-regain is comprised of multiple components (Fig. 1): time from birth to weight loss nadir (birth-tonadir), time from weight loss nadir to regain birthweight (nadir-to-regain), maximum weight loss percentage from birthweight, and a measure of degree of weight loss over time, or cumulative weight loss (calculated as the sum of the percent weight loss for each day the weight was below birthweight).Each of these components in uence the overall time-to-regain, and it is unclear what, if any, roles these factors have on neurodevelopmental outcomes.To examine the association of time-to-regain on outcomes, one must consider each component that impacts the overall time-to-regain.We performed a post-hoc, secondary analysis of the Preterm Erythropoietin (Epo) Neuroprotection Trial (PENUT) 20 to evaluate associations between time-to-regain, birth-to-nadir, nadir-to-regain, maximum weight loss percentage, and cumulative weight loss with neurodevelopmental outcomes at 2-years corrected age (CA).We hypothesized that rapid time-to-regain was associated with lower neurodevelopmental scores at 2-years CA.

Subjects and Methods
PENUT was a double-blinded, randomized, multi-center, placebo-controlled trial of EP newborns, de ned as 24w0d-27w6d at birth, evaluating Epo administration compared to placebo on neurodevelopmental outcomes. 20Details of the study protocol and primary outcomes have been previously published. 20All study subjects' legal guardians provided informed consent for their newborns to be included in PENUT.
Institutional Review Board approval was obtained at each site.
We conducted a secondary analysis of PENUT infants who survived to 2-years CA and underwent Bayley Scales of Infant and Toddler Development 3rd edition (Bayley-III) assessment.Daily uid and weight data over the rst 14 days were collected as part of the PENUT protocol.

Statistical Analyses
Associations between time-to-regain, including its four components, and each of the Bayley-III subscales were rst explored descriptively using locally estimated scatter plot smoothing (loess) curves.As weight data were not available beyond day 14, for visual representation and calculation of time-to-regain, any infant who had not regained birthweight by day 14 was assumed to have regained birthweight on day 15.We then determined quartiles of time-to-regain for further analyses (see Results).
For all inferential analyses, generalized estimating equations (GEE) with robust standard errors were used to appropriately account for potential correlation of outcomes for same-birth siblings. 21Baseline and demographic factors were compared across quartiles of time-to-regain using a multivariate Wald test.
Fluid administration over the rst 14 days was averaged relative to birthweight to provide a value of mL/kg birthweight/day.
Using GEE linear regression models adjusting for gestational age (GA) and Epo treatment group, we examined the association between neurodevelopmental outcomes at 2-years CA and maternal and infant factors, time-to-regain, birth-to-nadir, nadir-to-regain, maximum weight loss, or cumulative weight loss quartile.Infant factors that were clinically relevant and associated with faster time-to-regain (de ned as p-values ≤ 0.10) were then included as additional potential confounders when assessing the association between time-to-regain and neurodevelopmental outcomes. 21Maternal factors signi cantly associated with time-to-regain were not included as confounders in the nal outcomes models as they were all associated with birthweight, as detailed in the Results section (Table 1).Final regression models were adjusted for GA, treatment group, mode of delivery, vasopressor use in the rst 14 days (including dopamine, norepinephrine, epinephrine, dobutamine, vasopressin, or milrinone), postnatal steroid use in the rst 14 days (hydrocortisone or dexamethasone), birthweight z-score, maternal education, mechanical ventilation at baseline, and quartile of average total uids in the rst 14 days after birth.Relative to a reference quartile, we estimated the adjusted Bayley-III points associated with a given quartile of time-toregain, birth-to-nadir, nadir-to-regain, maximum weight loss, and cumulative weight loss.For time-toregain the reference quartile was de ned as ≥ 12 days, which is consistent with our hypothesis and also supported via an ancillary study among extremely low birthweight (ELBW) newborns. 5For birth-to-nadir, the reference quartile was de ned as 3-4 days, which is consistent with our previous publication demonstrating the majority of babies between 24-28 weeks' gestation reached their weight loss nadir at day 3. 16 For maximum weight loss, the reference quartile was de ned as -5 to -10% maximum weight loss, also consistent with our previous publication's ndings of weight loss of 5-15% being optimal for in-hospital outcomes. 16Similarly, when evaluating cumulative weight loss quartiles in adjusted models, -23 to < 0 percent-days was used as the reference group given our previous ndings suggesting that some weight loss (5-15%), but not severe weight loss (> 15%), is optimal for EP newborns. 16Therefore, we used the quartile that included some, but not severe, weight loss (-23 to < 0) as the reference category.7 (5-8)  7 (5-8)   7 (5-8)   7 (5-8)   7 (5-8 To determine the time-to-regain components most strongly associated with the outcomes of interest, a manual stepwise forward variable selection approach was used based on the Akaike information criterion (AIC) with a working independence model for multiples.AIC estimates prediction error, with a lower AICs generally indicating a better balance between predictive accuracy and model complexity.For each Bayley-III subscale, the AIC was determined for four separate adjusted models using quartiles of time-to-regain, birth-to-nadir, nadir-to-regain, maximum weight loss, and cumulative weight loss in addition to the covariates above.The model with the lowest AIC was initially selected, and the remaining four time-toregain components added individually in separate models for the next step.If one of these models had a lower AIC, then that model was selected, and the remaining three variables added in further models.This process continued until adding an additional time-to-regain component did not result in a lower AIC, and the model with the lowest AIC was selected for nal inference between the selected time-to-regain components adjusting for other variables as well as the pre-speci ed covariates.A p value < 0.05 was considered statistically signi cant.All analyses were conducted using the R statistical package (Version 3.6.1,Foundation for Statistical Computing, Vienna, Austria). 22

Association between time-to-regain components and 2-Year Neurodevelopmental Outcomes
Qualitatively, time-to-regain was positively associated with cognitive and motor Bayley-III scores (Supplemental Fig. 1A-C).When evaluating time-to-regain quartiles in adjusted models, newborns with the longest time-to regain (≥ 12 days) had higher mean Bayley-III cognitive, motor, and language scores compared to the lower three quartiles, but none were statistically signi cant (Supplemental Fig. 1D-E).There were no clear trends or statistically signi cant associations when evaluating birth-to-nadir, nadir-toregain, and maximum weight loss with any Bayley-III score at 2-years CA in adjusted models, although nadir-to-regain approached signi cance with more rapid (≤ 1 day) nadir-to-regain compared to ≥ 8 days (Supplemental Figs.2-4).

Stepwise selection of multiple weight loss variables
After adjusting for GA, treatment group, vasopressors, steroid use, small for gestational age status (SGA), uid quartiles, mechanical ventilation, and maternal education, the models including cumulative weight loss quartiles had the lowest AIC for predicting motor and language scores, whereas the nadir-to-regain quartile model had the lowest AIC for predicting motor scores.Adding cumulative weight loss quartiles further lowered the AIC for predicting motor scores, with the addition of nadir-to-regain quartiles lowering the cognitive and language score predictions.The addition of birth-to-nadir, maximum weight loss, and/or time-to-regain quartiles to these models did not further lower AIC.Therefore, the nal combined models included both nadir-to-regain and cumulative weight loss to predict all Bayley-III subscales.In the nal fully adjusted models (Table 2), signi cantly lower cognitive scores were seen in those with higher overall cumulative weight loss (≤-50 percent-days: -5.0 Bayley III points, 95% CI -8.9, -1.1 compared with − 23 to 0 percent-days) or those with a rapid nadir-to-regain (≤ 1 day: -5.3 Bayley III points, 95% CI -9.8, -0.8 compared with ≥ 8 days).Also, increasing cumulative weight loss was also associated with lower Bayley-III language scores (≤-50 percent-days: -7.2 Bayley III points, 95% CI -11.8, -2.6 compared with − 23 to 0 percent-days).

Discussion
This study examined the association between time to regain birthweight and its components including birth-to-nadir, nadir-to-regain, maximum weight loss, and cumulative weight loss with 2-year neurodevelopmental outcomes.While overall time-to-regain quartile was not statistically associated with neurodevelopmental outcomes, a faster nadir-to-regain and increasing cumulative weight loss were associated with lower Bayley-III cognitive scores in the nal, AIC, stepwise forward regression model.These associations were more apparent once both cumulative weight loss and nadir-to-regain were included together in the nal model, con rming that knowledge of multiple components of weight loss may be required to fully understand their potential impact on outcomes.For instance, nadir-to-regain was not associated with outcomes when examined on its own, but a signi cant association between rapid nadir-to-regain and adverse neurodevelopmental outcomes was seen once also adjusting for cumulative weight loss.Ultimately, these ndings suggest that too rapid a recovery of weight loss (faster nadir-toregain) or prolonged, excessive cumulative weight loss are potentially detrimental to long-term outcomes of EP newborns.These results suggest a balance of preventing severe, prolonged weight loss while ensuring birthweight is not regained rapidly after the weight loss among EP newborns.
We have not found other publications exploring time-to-regain and its components on long-term outcomes of EP newborns.However, there are several studies that suggest a longer time-to-regain is associated with adverse in-hospital health outcomes among preterm newborns.Speci cally, in a singlecenter, retrospective case-control study in Australia, preterm newborns with type I retinopathy of prematurity had a longer time-to-regain when compared to their matched controls, but these ndings did not reach statistical signi cance (median day of time-to-regain 9 vs 7, adjusted odds ratio 1.08, 95% CI 1.00-1.17,p = 0.059). 6The authors concluded that these ndings suggest that a longer time-to-regain may aid clinicians in predicting which newborns have retinopathy of prematurity. 6Similarly, indicators for malnutrition for preterm infants and neonates were developed by neonatal dietitians who note that longer time-to-regain (> 15 days) among preterm newborns is associated with a diagnosis of malnutrition. 19aken together, current recommendations and published ndings suggest a shorter time-to-regain is optimal for preterm newborn outcomes.However, these studies and recommendations do not include evaluation of the rapidity of time-to-regain and its associated components which likely in uence overall time-to-regain.Furthermore, these studies do not assess their association with neurodevelopmental outcomes, nor are they speci cally focused on the EP population.
Similar to our results, another study reported a longer time-to-regain may be optimal for long-term outcomes among ELBW newborns.Ehrenkranz and colleagues sought to evaluate the association between in-hospital growth velocity and neurodevelopmental outcomes at 18-22 months' CA among 495 newborns with birthweights between 501-1000 grams. 5They divided their cohort into quartiles based upon average in-hospital growth velocities in which growth velocity was de ned as "the period between the time that the infant regained birth weight and discharge, transfer, age 120 days, or until a body weight of 2000g was reached." 5 The lowest quartile for growth velocity included newborns with a mean weight gain of 12.0 g/kg/day whereas the highest quartile included newborns with a mean weight gain of 21.2 g/kg/day. 5Newborns in this highest quartile for growth velocity had signi cantly less neurodevelopmental impairment compared to the lowest quartile (29% vs 55%, p < 0.001). 5Interestingly, newborns in the highest quartile for growth velocity, the group with the least neurodevelopmental impairment, had longer time-to-regain than newborns in the lowest quartile (19.5 days vs 15.9 days, p = 0.003).5 All 4 quartiles had average time-to-regain more than 14 days after birth, supporting that ELBW newborns commonly require more than 2 weeks to regain birthweight. 5he current study is novel because it focuses on weight parameters in the rst 14 days and their association with neurodevelopment.Most studies that assess the association of neonatal growth velocity on later neurodevelopment outcomes generally either do not include the different components of time-to-regain in the analyses or start after birthweight has been regained, thereby overlooking the critical fetal-to-neonatal transition. 5,23,24Our study's ndings highlight the need for further exploration into the ideal weight loss patterns among EP newborns with respect to long-term outcomes through prospective, randomized trials.

Biologic Associations with Time to Birthweight
It is uncertain what constitutes optimal rate of lean body mass growth velocity during the rst 2 weeks after birth among EP newborns.In fact, it is likely that time-to-regain and its components are more a marker of changes in uid status rather than alterations in lean body mass.][9][10][11][12][13][14][15][16] These factors collectively predispose EP newborns to extraordinary uid and weight loss in the immediate postnatal period.Initial postnatal physiologic weight loss is a necessary, important factor that is associated with improved health outcomes of preterm newborns.This process of physiologic weight loss after birth is critical to a proper fetal-to-neonatal transition, and a lack of weight loss is associated with adverse outcomes including BPD and NEC, among others. 25,16erall, the fetal-to-neonatal transition that occurs over the rst weeks after birth is a complex period that involves the interplay between metabolic, endocrine, and uid alterations.In congruence with other publications demonstrating that minimal or a lack of weight loss after birth is associated with adverse health outcomes, 16,25 our ndings similarly demonstrate adverse neurodevelopmental associations with more rapid time-to-regain (speci cally, a more rapid nadir-to-regain) and prolonged, excessive weight loss (increasingly negative cumulative weight loss) among EP newborns.

Strengths and Limitations
The of our study include the relatively large sample size of 654 EP neonates from 19 academic centers including 30 NICUs across 13 states in the United States. 20Due to the multi-center nature of this study, the results may be generalizable to other U.S.-based NICUs and other higher-resourced settings globally.Our study also used a contemporary cohort describing the association between time-to-regain and 2-year neurodevelopmental outcomes among EP infants.
While we attempted to control for variables associated with increased severity of illness such as increased total uid administration over the rst 14 days, postnatal steroid use, postnatal vasopressor use, lower birthweight, and need for mechanical ventilation, it is possible that rapid time-to-regain and associated components are markers of increased severity of illness.Our analyses cannot delineate when or if time-to-regain and associated components re ect uid management changes versus lean body mass development.Moreover, given that our study occurred in U.S.-based institutions, our ndings may not be generalizable to low-and middle-income settings where access to humidi cation, incubators, intravenous uids, and other resources may in uence both short-and long-term health outcomes.As PENUT only included daily weights through the rst 14 days after birth, we are unable to determine the exact timing of regaining birthweight beyond 14 days.This lack of data prevents determination of a potential optimal window for time-to-regain.Finally, our study is a retrospective, secondary analysis of the PENUT trial, and, as such, cannot attribute causality.Future studies targeting speci c weight loss ranges or trajectories should be further tested in randomized controlled trials to discern causality prior to implementing these ndings in clinical practice.

Conclusion
Our Con icts Interest: The authors report no con icts of interest.
Author Contributions: GCV and conceptualized the design of the secondary analysis, composed the initial draft of the manuscript, and revised its subsequent versions.TRW performed the statistical analyses.TRW, BAC, DEM, SK, KMS, OCB, KMS, JBL, PJH and SEJ were involved in revisions to the manuscript, and all agreed to the nal draft of the manuscript being submitted.

Data Sharing Statement:
De-identi ed individual participant data is available through the NINDSData Archive: https://www-nindsnih-gov.offcampus.lib.washington.edu/Current-Research/Research-Funded-NINDS/Clinical-Research/Archived-Clinical-Research-Datasets.The data is de-identi ed and a limited access data set is available through a request form on that page.Data dictionaries, in addition to study protocol, the statistical analysis plan, and the informed consent form will be included.The data will be made available upon publication of all PENUT Trial related manuscripts to researchers who provide a methodologically sound proposal for use in achieving the goals of the approved proposal.
Con icts of Interest: All authors declare they have no nancial interests or con icts of interest in relation to the work described in this manuscript.

Supplementary Files
This is a list of supplementary les associated with this preprint.Click to download. TRBSupplementalMaterialsJperi82.docx

Figures
Figures

Figure 1 Time
Figure 1

Figure 2 A
Figure 2

Table 1
Maternal and infant demographic, baseline, and exposure variables by time-to-regain quartile.Categorical variables represented by n (%).Continuous variables represented by mean (SD).
*P-value compares differences across weight regain groups using a multivariate Wald test adjusting for gestational age and Epo treatment group.

Table 2
Stepwise forward AIC fully-adjusted model demonstrating signi cant associations between cumulative weight loss and nadir-to-regain with Bayley-III 2-year CA neurodevelopmental outcomes.Adjusted for gestational age, treatment group (Epo or control), vasopressors or steroids use in the rst 14 days after birth, birthweight z score, uid quartiles, maternal education, and mechanical ventilation.Birth-to-nadir, time-to-regain, and maximum weight loss were not selected as predictors in the nal model.
study highlights that rapid regain of birthweight, especially rapid regain from the weight loss nadir, and prolonged, severe weight loss, are associated with adverse neurodevelopmental outcomes among EP newborns at 2 years CA.Our ndings suggest that clinicians caring for EP newborns should consider allowing moderate physiologic weight loss with subsequent gradual time to regain birthweight.PENUT was supported by the National Institute of Neurological Disorders and Stroke of the National Institutes of Health under award numbers U01NS077953 and U01NS077955.