As this is an observational study, there is no intervention involved with this protocol. Many study visits for the Neuroimaging Ancillary Study coincide with visits included in the RAPIDIRON or RAPIDIRON-KIDS protocols. Protocol details for these trials have been previously published ( Derman et al. 2021, 2023). Participation in this study can begin at any point during the third trimester of pregnancy up to 4 months 15 days post-delivery. All participants will be followed through 24 months of age of the offspring. Thus, participants recruited during pregnancy will have a duration of participation of approximately 26 months, while those recruited post-delivery will participate for approximately 20-24 months.

The Neuroimaging Ancillary study will be conducted in Belagavi, India with participants recruited from the RAPIDIRON and RAPIDIRON-KIDS. Study visits will occur at a single hospital— KLES Dr. Prabhakar Kore Charity Hospital and Medical Research Centre in Belagavi, which has a 3 Tesla equipment required for performing MRI (3 Tesla Magetom Spectra) as well as data collection capabilities. Transportation will be made available to participants to facilitate attendance of study assessments.

For this ancillary study, only mothers and infants of randomized RAPIDIRON and RAPIDIRON-KIDS participants recruited in Karnataka will be eligible for participation. Additional eligibility criteria will include:

Inclusion criteria : •

Maternal participant already in the RAPIDIRON Trial – currently in third trimester of pregnancy or <4 months 15 days post-delivery.

Exclusion criteria: •

Contraindications for undergoing MRI scan like those with metallic or other surgical implants, pacemaker, heart valves, aneurysm clips, metal plates or pins and some orthopedic prostheses and metal pieces anywhere in the body.

Community sensitization efforts were conducted in areas served by the primary and community health centers involved in the RAPIDIRON Trial. Additional sensitization will take place in the Karnataka research area to raise awareness about the RAPIDIRON-KIDS study and this Neuroimaging Ancillary Study. These efforts typically involve research staff meeting with community leaders, health officials, and other relevant stakeholders. Accredited Social Health Activists (ASHAs) residing in the research communities may be invited to participate in the study, with authorization from their supervising health official. Their roles may include accompanying mothers and children to engagement centers for screening and consent processes or ensuring that participants attend appointments for blood draws or developmental testing. ASHAs receive incentive-based fees for their work, which includes transporting mothers and children to health facilities for necessary care. Their involvement has been crucial in previous research aimed at improving women’s and children’s health. ASHAs with education on anemia and iron deficiency can also provide additional information about this study.

Current RAPIDIRON Trial participants enrolled in the RAPIDIRON-KIDS follow-up study at the Belagavi site in Karnataka will receive information about this trial. Interested participants will be screened and given the opportunity to provide consent from the third trimester up to four months post-delivery. Before consenting, potential participants will have ample time to consider participation and discuss with trusted family and friends before making a decision regarding participation. They also have the opportunity to discuss procedures and ask questions of the study staff. The consent form is available in multiple local languages, and the consent process will include a verbal review to accommodate varying literacy levels. Informed written consent will be obtained by a physician or other certified healthcare provider. Informed consent of the RAPIDIRON participant is obtained for her participation and that of her offspring in the ancillary trial.

This is a prospective longitudinal study with assessments up to four time points across early development: prenatal (third trimester), 4, 12, and 24 months ( Table 1). Each time point will include collection of magnetic resonance imaging (MRI) for neuroimaging, child health (e.g. child length, weight, and head circumference) measurements, and child cognition and behavioral assessments. Additional biospecimen collection and auditory brain response acquisition is set to be performed at the 4-month visit. Additional details of the study protocol and procedures are described below.

Fetal MRI has increasingly become a safe and valuable tool for assessing fetal brain development in utero. With superior soft tissue contrast compared to ultrasound, fetal MRI enables differentiation of developing brain structures and tissues, including the cerebral cortex, deep gray matter, white matter, and cerebrospinal fluid (CSF) spaces ( Peña-Trujillo et al. 2024). Moreover, advances in faster MRI sequences to reduce image acquisition times as well as the development of motion-robust image reconstruction and segmentation tools have allowed fetal MRI to be utilized for quantitative assessment.

Pregnant participants recruited prior to delivery will undergo fetal magnetic resonance imaging (MRI) to assess brain development in utero. We understand that MRI scan sessions can be challenging given the stage of pregnancy during which we are scanning, thus, we have limited our acquisition protocol to include structural and diffusion imaging metrics that have high reliability. Scan sessions are limited to approximately 40 minutes (including set up) and, if needed, breaks are offered throughout the scan to allow mothers to roll or change positions. There will be an outpatient room made available for the participant for their comfort and convenience.

MRI scans will be performed using the 3T Siemens Magetom Spectra 3T scanner with sequences optimized for fetal imaging. The imaging protocol consists of structural T2-weighted and diffusion-weighted imaging to provide comprehensive anatomical and microstructural data. T2-weighted images are acquired using the rapid two-dimensional Half Fourier Single-shot Turbo spin-Echo (HASTE) sequence ( Semelka et al. 1996) while diffusion weighted images are collected with a single-shot spin-echo echo planar imaging (EPI) sequence. Additional fetal MRI sequence parameters are included in Table 2. Structural and diffusion images are repeated in multiple orthogonal planes (transverse, coronal, sagittal) to allow for slice-to-volume reconstruction ( Figure 1) ( Uus et al. 2020; Kuklisova-Murgasova et al. 2012).

For children participating in this study, MRI scans are performed during natural, non-sedated sleep ( Dean et al. 2014; Raschle et al. 2009, 2012; Spann et al. 2023). This approach leverages strategies pioneered by team members and adopted globally, including night-time scanning sessions and the use of MRI-safe immobilizers (CFI Medical Solutions, Fenton, MI, USA) that help keep the children warm and minimize movement during the scan, ensuring high-quality imaging. MRI sessions are scheduled around parents’ schedules and the child’s sleep patterns (e.g. nap and bedtimes). During the MRI scan session, a trained research team member will continuously monitor the sleeping child within the scan suite to ensure they remain asleep and comfortable. Additionally, the child’s mother and/or father are welcome to stay in the scanner suite, to offer reassurance and support provided they have no MRI contraindications. To further enhance comfort and convenience to participants, an outpatient room will be available overnight. This room is equipped to provide a restful environment for both the child and their parents, allowing them to relax before and after the scan.

All imaging is performed using the same 3T Siemens Magetom Spectra 3T scanner using an 8-channel radio-frequency imaging head coil for all visits. A comprehensive brain imaging protocol was developed to provide detailed information regarding brain morphology and microstructure, including structural, diffusion, and quantitative relaxometry. Detailed description of the imaging protocol is below. Total acquisition time is approximately an 40 minutes. While the goal is to collect as much of the scan protocol for all children, it is recognized that this can be a challenge for pediatric neuroimaging ( Spann et al. 2023; Dean et al. 2014). Thus, a minimum acquisition of the structural, diffusion and quantitative T2 relaxometry (see below) was established as criteria for a successful acquisition. Additional images were acquired permitting the child remained asleep throughout the duration of the scan.

Structural Imaging: Postnatal T1-weighted (T1w) and T2-weighted (T2w) imaging will be acquired to provide information on brain morphometry and structural development, allowing measurements of whole-brain, regional and cortical volumes, thickness, and surface area. T1w imaging will be obtained using the three dimensional magnetization–prepared rapid gradient echo imaging (MPRAGE) ( Mugler and Brookeman 1990), while T2w imaging will be acquired using the three-dimensional variable flip angle turbo spin echo sequence (T2-SPACE). Additional scan parameters are detailed in Table 2.

Diffusion Weighted Imaging: Diffusion MRI, and more specifically diffusion tensor imaging (DTI), has become one of the most widely used techniques for characterizing brain tissue microstructure, organization, and structural brain connectivity ( refs). Diffusion MRI will be acquired with the product diffusion weighted echo planar imaging sequence. Two separate acquisitions of diffusion weighted images will be performed, one with 12 and one with 6 non-collinear diffusion encoding directions with a b = 1000 s/mm ² and a single non-diffusion weighted (e.g. b = 0 s/mm ²) image. The acquisition is repeated two times for a total acquisition time of approximately 6.5 minutes. Additional parameters are detailed in Table 2.

Quantitative Relaxometry Measurement of the intrinsic relaxation times (e.g. T ₁ and T ₂), commonly referred as relaxometry, can facilitate improved characterization of brain tissue, enhance image contrast, and provide a more direct link and interpretation between observed intensity changes ( Deoni 2010). Moreover, given the dependence of T ₁ and T ₂ on the local tissue environment, including water content, macromolecules, protein and lipid composition, and paramagnetic atom (e.g iron) concentration, changes in T ₁ and T ₂ can be indicative of alterations linked to disease, pathology, or other biological processes ( Deoni 2010). A multiple echo-time spin-echo strategy was implemented as part of the MRI protocol to measure T ₂ ( Poon and Henkelman 1992), with three separate echo times (TE = 16, 79, 141 ms) acquired. For measurement T ₁, the rapid and time-efficient driven equilibrium single pulse observation of T1 (DESPOT1), or variable flip angle, strategy was developed ( Deoni et al. 2003; Christensen et al. 1974) with three spoiled gradient echo images acquired with flip angles of α = 4°, 10°, 20°. To correct for inhomogeneity of the transmit (B1+) field, a B1 field map was additionally included as part of the protocol. See Table 2 for additional scan parameters.

Additional Imaging Sequences In addition to the techniques described above, additional methods were implemented as part of the protocol to maximize the information obtained. For example, a magnetization transfer weighted image will be acquired and be used to provide information regarding the macromolecular content of the brain ( Sled 2018). A three-dimensional arterial spin label (ASL) acquisition was additionally implemented to assess cerebral perfusion ( Petcharunpaisan et al. 2010). Additional scan parameters are provided in Table 2.

The parent RAPIDIRON Trial involves blood draws during pregnancy and cord blood collection at birth to measure hemoglobin (Hb) status in both the pregnant participant and the newborn. Additionally, the RAPIDIRON-KIDS protocol includes collecting whole blood samples from the mother at 4, 12, and 24 months post-delivery, and from the child via venipuncture at 4 months and heel stick at 12 months, to assess hemoglobin and iron levels. These data will also be utilized for analyses in the Neuroimaging Ancillary Study.

Additionally, in this study, nail samples from the mother during pregnancy and from both the mother and child at the 4-month visit will be obtained to measure cortisol levels, an indicator of chronic stress.

Birth weight and length will be measured by the study team as part of the RAPIDIRON Trial. Information on labor and delivery practices, as well as any complications, will be gathered from maternal reports. These data will help identify children who experience significant delivery trauma or complications that may have neurological implications. Additionally, child length, weight, and head circumference will be measured at 4, 12, and 24 months of age to assess the incidence of stunting and wasting.

Auditory brainstem evoked response (ABR) is a non-invasive neurophysiologic assessment of auditory neural maturation, a surrogate outcome for brain maturation and especially myelination in premature infants. The ABR waveform is comprised of 3 waves (I, III, and V). Wave I is generated peripherally in the auditory nerve. Wave III reflects the firing of axons exiting the cochlear nuclear complex in the brainstem, and wave V primarily reflects an action potential generated by axons from the lateral lemniscus at a more rostral brainstem location. ABRs are recorded with a standard system used in audiology clinics with the subject lying supine. Testing should be performed by audiologists skilled in the administration of ABR tests. Electrode sites are mastoid (reference), midline on high forehead or crown of the head (active), and shoulder (ground). Electrode gel is applied to silver/silver chloride electrodes. Bilateral monaural ABR tests are performed using 80 dB nHL broadband click stimuli with insert earphones. The clicks are typically presented at a repetition rate of 29.9/sec, and three runs of 2000 repetitions are recorded for each ear ( Young et al. 2025).

This assessment will be conducted at 4 months of age for all offspring participants. The study audiologist will perform the test, and the concerned data form will be completed by the study pediatric neurologist on a weekly basis.

As part of the RAPIDIRON-KIDS Study, infants will undergo a series of neurobehavioral assessments at 4, 12, and 24 months of age. These assessments will include both general developmental tests and specific evaluations targeting brain regions likely affected by iron status. The tests administered will be the Ages & Stages Questionnaire-3 ( Squires et al. 2009) ( ref), Preferential Looking Time ( Fagan 1974), Bayley Scales of Infant Development-IV (a primary outcome) ( Bayley and Aylward 2019), and the Behavioral Rating Scale ( Lozoff et al. 1998). Data derived from the parent trial(s) will be analyzed to explore associations between neurocognitive outcomes and MRI-based metrics in this ancillary study.

Maternal depression will be assessed using the Edinburgh Postnatal Depression Scale (EPDS) ( Cox et al. 1987; Adouard et al. 2005), which has been validated for use in Indian populations ( Russell et al. 2020) at prenatal, 4-, 12-, and 24-month visits. Additionally, maternal perceived stress will be measured with the Perceived Stress Scale (PSS-10) ( Cohen et al. 1983) at prenatal, 4-, and 24-month visits. Participants recruited post-delivery will complete these assessments at 4, 12, and 24 months. As part of the RAPIDIRON-KIDS Study, maternal quality of life will be assessed using the World Health Organization Disability Assessment Schedule 2.0 (WHODAS 2.0) ( Konecky et al. 2014) at 4 and 12 months.

Data will be collected primarily with the use of paper-based forms. JNMC has successfully used the Research Electronic Data Capture (REDCap) system for numerous research projects. This system is a secure web application for building and managing online surveys and databases, and it offers a secure, privacy compliant web-based electronic data capture and maintenance system. The data are stored in a centralized system. REDCap ( https://project-redcap.org) software requires each user to have their own account, and user privileges are applied to ensure that users have access only to data and information they need and are authorized to have.

Flywheel ( https://flywheel.io), a medical imaging data management and analysis platform, is used to streamline imaging data management and to automate research workflows. Flywheel offers a unified and secure platform for data storage, curation, and analysis. This platform enables the development, testing, and execution of harmonized analysis pipelines across imaging datasets, with comprehensive tracking of data provenance. Importantly, this can be achieved without the need for additional local hardware infrastructure or software installations. Flywheel has been deployed in India to address the data-sharing restrictions in India and is used to host imaging data acquired in the Neuroimaging Ancillary Study.

The primary analysis will involve comparisons of the combined IV iron group to the oral iron group with respect to three outcomes from the fetal and 4-month MRI scans. The hippocampus volume and FA and MD at the posterior limb of internal capsule will be a primary target as these are early developing brain regions. We will estimate group differences at all measurement times with the fetal and 4-month comparisons being the main tests of interest. Each comparison will be performed at the alpha=0.01 level. We anticipate 15 fetal scans from mothers who received oral iron and 30 from mothers who received IV iron giving 80% power to detect a standardized effect size of 1.12 standard deviations. For the 4-month comparisons, with expected sample size of 55 subjects whose mothers received oral iron and 110 subjects whose mothers received IV iron, we have 80% power to detect a standardized effect size of 0.57 standard deviations. From published data on scans within weeks of birth (ref ), the standard deviation of side-specific hippocampal volumes ranges from 0.23-0.29 cc. Thus, we are powered to detect a mean group difference of 0.17 cc at 4 months (when the SD is 0.29). The standard deviation of FA and MD values at 1 month are in the range of 0.2-0.3 ( Dean et al. 2017). For these outcomes, the detectable difference at 4 months in means is between 0.11 and 0.17. These are conservative power calculations as they ignore the fact that we will have 2 measures of each outcome (1 per side) for each subject and ignore potential borrowing of information across the multiple longitudinal outcomes. For each outcome, we will use mixed effects linear regression to jointly model both measurements including group (IV vs. oral), side, measurement time (fetal, 4, 12, 24 month), and group by time interaction as fixed effects and an appropriate residual covariance matrix structure to account for correlation among measurements in the same subject at the same time and across time.

Comparisons of secondary fetal and infant imaging outcomes will use a similar approach to the primary data analysis. Each outcome will be modeled using a linear mixed effects regression model to include time and treating time of measurement as a fixed effect to allow for maximum flexibility in trajectory shape. An interaction of group and time will allow for exploring differences in change of imaging parameters over time by mother’s iron treatment.

We will also consider dimension reduction techniques such as principal component analysis (including sparse versions) so that time dynamics can be summarized by one, or a few, trajectories. We will then model the summary measures using the same longitudinal mixed effects approach.

Other secondary outcome measures (e.g., maternal and infant cortisol measures, maternal mental health and well-being) will be analyzed using linear regression for outcomes measured at one time point or linear mixed effects regression for outcomes measured longitudinally.

CTRI Registration: CTRI/2022/11/047102

The study protocol, informed consent forms and any protocol modifications, have been reviewed and approved by the Thomas Jefferson University IRB (09/08/2022; iRISID-2022-0823), and the Institutional Ethics Committee, KLE Academy of Higher Education and Research, Belagavi, India (09/30/2022; KAHER/EC/2022-23/D-19102205).

Informed written consent is obtained from the RAPIDIRON participant for her participation and the participation of her offspring in the Neuroimaging ancillary trial.