Rapid normalization of vitamin D deficiency in PICU (VITdALIZE-KIDS): study protocol for a phase III, multicenter randomized controlled trial

Abstract

Background

The rate of vitamin D deficiency (VDD) in critically ill children worldwide has been estimated at 50%. These children are at risk of multiple organ dysfunction, chronic morbidity, and decreased health related quality of life (HRQL). Pediatric and adult ICU clinical trials suggest that VDD is associated with worse clinical outcomes, although data from supplementation trials are limited and inconclusive. Our group’s phase II multicenter dose evaluation pilot study established the efficacy and safety of an enteral weight-based cholecalciferol loading dose to rapidly restore vitamin D levels in critically ill children.

Methods

Our aim is to evaluate the impact of this dosing regimen on clinical outcomes. VITdALIZE-KIDS is a pragmatic, phase III, multicenter, double-blind RCT aiming to randomize 766 critically ill children from Canadian PICUs. Participants are randomized using a 1:1 scheme to receive a single dose at enrollment of enteral cholecalciferol (10,000 IU/kg, max 400,000 IU) or placebo. Eligibility criteria include critically ill children aged newborn (> 37 weeks corrected gestational age) to < 18 years who have blood total 25-hydroxyvitamin D < 50 nmol/L. The primary objective is to determine if rapid normalization of vitamin D status improves HRQL at 28 days following enrollment. The secondary objective is to evaluate the impact of rapid normalization of vitamin D status on multiple organ dysfunction. The study includes additional tertiary outcomes including functional status, HRQL and mortality at hospital discharge and 90 days, PICU and hospital length of stay, and adverse events related to vitamin D toxicity. Additionally, we are performing comprehensive vitamin D speciation and non-targeted metabolite profiling as part of a sub-study for the first 100 participants from whom an enrollment and at least one post-intervention blood and urine sample were obtained.

Discussion

The VITdALIZE-KIDS trial is the first phase III, multicenter trial to evaluate whether rapid normalization of vitamin D status could represent a simple, inexpensive, and safe means of improving outcomes following pediatric critical illness. Recruitment was initiated in June 2019 and is expected to continue to March 2026.

Trial registration

Clinicaltrials.gov, NCT03742505. Study first submitted on November 12, 2018 https://clinicaltrials.gov/study/NCT03742505

Background

Approximately 12,000 critically ill children are admitted to Canadian pediatric intensive care units (PICUs) annually, where they face the risk of prolonged rehabilitation, new morbidity or chronic disease, and death [1]. Research has shown that a significant number of PICU survivors experience a long-term decline in functional status and/or health-related quality of life (HRQL) [2,3,4]. Novel means of decreasing mortality and reducing long-term morbidity would be of great value to these children, their families, and the healthcare system.

Given vitamin D’s established pleiotropic role in the health of organs central to critical illness pathophysiology and support of immune function, it has been hypothesized that vitamin D deficiency (VDD) could represent a modifiable risk factor for improving clinical outcome following critical illness [5, 6]. Specifically, there are multiple mechanisms through which VDD could contribute to organ dysfunction, including (i) exacerbation of critical illness-related hypocalcemia (particularly in the presence of liver, parathyroid, or renal dysfunction) [7]; (ii) cardiovascular dysfunction indirectly through low body calcium stores and directly through vitamin D receptors (VDR) present on myocytes and endothelial cells; (iii) immune dysregulation through functional VDR present on all major immune cell types [8]; (iv) through the role of vitamin D signaling in innate immunity [9]; and (v) muscle weakness, a well-recognized consequence of critical illness [10] that can be further exacerbated by the impact of VDD on muscle pathology and clinically relevant weakness [11].

Multiple observational studies in the PICU have reported both high deficiency rates and associations between VDD and organ dysfunction, health resource utilization, and mortality [12,13,14,15,16]. The prevalence and impact of VDD in critically ill children is well established, with meta-analysis of observational studies showing a worldwide PICU VDD rate of 50% [16]. Our research group also published the first large Canadian PICU study (n = 326) documenting high deficiency rates (60–70%) [12, 17] and association between lower vitamin D levels and worse clinical course in the PICU setting [12]. Although concerning, the high VDD rate in critically ill Canadian children is viewed by our research group and others [18,19,20] as an opportunity to potentially improve clinical outcomes, with far-reaching implications for children, their families, and the healthcare system. The gap in knowledge at present is whether early rapid normalization of VDD in critically ill children will causally improve outcomes.

Similar to pediatric critical illness, there is a significant body of observational evidence demonstrating VDD to be risk factor in the adult ICU setting [21,22,23,24,25]. In contrast to pediatrics, there have been ~ 10 RCTs evaluating the biochemical and clinical response to high dose vitamin D in critically ill adults. Meta-analyses of trial findings are inconclusive as to whether there is benefit [23, 24, 26, 27]. The definitive adult ICU trial (VITdALIZE), a multi-national RCT seeking to enroll 2400 critically ill VDD adults to determine whether rapid normalization with a 540,000 IU load improves 28-day mortality, is ongoing [28].

Current vitamin D supplementation guidelines

Vitamin D supplementation to achieve recommended daily allowances of intake thought essential to good health has been recommended in guidelines from multiple agencies. For example, the National Academy of Medicine (NAM, formerly the Institute of Medicine) provides a recommended daily allowance or adequate intake (400–600 IU/day) for children [29] that has been supported by Health Canada. As there are no ICU guidelines recommending alternative loading dose supplementation strategies, most critically ill children receive either no supplementation or the same daily dose suggested for healthy children. It is important to recognize these doses (i) are derived from studies on healthy children, (ii) are focused on preventing bone disease, and (iii) may require months to rebuild body stores [30]. Emphasizing this last point, evaluation of hospitalized patients in observational studies and clinical trials has shown that vitamin D levels often fail to rise or even fall with usual care [13, 22, 31].

In addition to standard dosing, the NAM also provides alternative dosing recommendations, referred to as the tolerable upper intake level (UL) for healthy children, ranging from 1000 to 4000 IU/day (age dependent) [29]. We recently investigated this daily dosing regimen for potential application in the PICU setting in a systematic review [32], demonstrating that the majority of VDD patients require over a month to attain normal levels at the UL dose. This time frame was confirmed in a recent pilot study of pediatric burn patients, with only 50% of patients achieving normal levels despite 30 days of daily dosing above the UL (~ 800 IU + 100 IU/kg) [33]. The use of a supplementation approach that requires a month (or more) to correct VDD may not optimize benefit during the PICU admission and immediate recovery period, placing critically ill children at risk for new pathology and worse functional outcome during these times [33]. While this dosing regimen is of value for certain patient populations in the context of early treatment or prophylactic applications, it offers minimal value for acutely critically ill children.

Pediatric investigations to date

A novel supplementation approach that rapidly normalizes vitamin D levels is essential to realize the potential health benefits of sufficient vitamin D status. To determine the loading dose most appropriate for a randomized controlled trial evaluating rapid correction of VDD in critically ill children, we analyzed vitamin D response in 98 pediatric clinical trials (88 full-text publications, 10 conference abstracts) [32]. This systematic review demonstrated that administration of a single or repeated enteral loading dose rapidly achieved peak vitamin D (measured as blood total 25OHD) levels within 3 days of dosing [32]. Further analysis showed increased hypercalcemia risk with loading doses > 400,000 IU, but only in studies enrolling infants and young children. Based on this review, a single load of 10,000 IU/kg (maximum 400,000 IU) was determined as the regimen most likely to rapidly and safely correct VDD [32].

We subsequently evaluated the safety and efficacy of a 10,000 IU/kg (maximum 400,000 IU) loading dose in our international phase II dose evaluation pilot study VITdAL-PICU. Results from this trial confirmed this dose raised group 25OHD levels to above the 75 nmol/L threshold in > 75% of participants, with no biochemical or clinical evidence of toxicity [34]. Similarly, in a trial of VDD children undergoing CHD cardiac surgery, a loading dose of 10,000 IU/kg cholecalciferol (maximum 400,000 IU) approximately 2 weeks prior to surgery produced significantly higher concentrations in the treatment arm relative to control (83.5 nmol/L vs 27.4 nmol/L) in blood collected immediately before the surgery, without statistically significant differences in parameters pertaining to vitamin D toxicity during the perioperative period [35]. Recent RCTs in critically ill pediatric populations also support findings of negative impact of VDD on clinical outcomes for sepsis and pneumonia and suggest a positive effect of loading dose supplementation in these populations [36,37,38].

Trial outline

VITdALIZE-KIDS will investigate the hypothesis that loading dose vitamin D could represent a simple, inexpensive, safe, and feasible means of improving outcomes following pediatric critical illness. This trial asks the question: in critically ill VDD children, what is the effect of rapid normalization of vitamin D status compared with placebo, on health-related quality of life (HRQL) at 28 days following intervention? The secondary objective is to determine the impact of rapid normalization of vitamin D status on new and progressive multiple organ dysfunction. Tertiary objectives are to evaluate the impact of rapid normalization of vitamin D status on (1) functional status at hospital discharge and 28 and 90 days following loading dose administration; (2) PICU and hospital length of stay; (3) serious adverse events and vitamin D-related toxicities; (4) survival alone at 28 and 90 days; and (5) HRQL at hospital discharge and 90 days following loading dose administration. Further, VITdALIZE-KIDS includes a sub-study of 100 participants with pre- and post-loading dose blood and urine, where we will assess pre- and post-intervention vitamin D status, perform non-targeted metabolite profiling to better understand the pleiotropic effects of rapid correction of VDD on human metabolism beyond calcium homeostasis and active hormone levels, and identify adjunct biomarkers that when combined with 25OHD may better predict inadequate vitamin D status and response to treatment.

Methods/design

In collaboration with the Canadian Critical Care Trials Group (CCCTG), we designed VITdALIZE-KIDS as a pragmatic, phase III, multicenter, individual patient 1:1 randomized double-blind superiority RCT. Our protocol adheres to the Standard Protocol Items: Recommendations for Interventional Trials (SPIRIT) checklist (see Additional file 1). A summary of the protocol is provided in Table 1.

Table 1 World Health Organization trial registration data set—structured summary

Trial setting

The VITdALIZE-KIDS trial will recruit up to 766 patients from PICUs at the following 11 academic centers in Canada: Children’s Hospital of Eastern Ontario (CHEO), Ottawa; London Health Sciences Centre, London; Montreal Children’s Hospital, Montreal; Centre hospitalier universitaire Sainte-Justine, Montreal; McMaster Children’s Hospital, Hamilton; Alberta Children’s Hospital, Calgary; BC Children’s Hospital, Vancouver; The Hospital for Sick Children, Toronto; Stollery Children’s Hospital, Edmonton; IWK Health Centre, Halifax; and Centre mère-enfant Soleil du CHU de Québec Université Laval, Quebec City.

Patient enrollment and consent

Eligible participants are identified in the PICUs. Patients are screened daily (Monday through Friday) by trial staff and informed consent obtained via site-approved approaches, according to Good Clinical Practice (GCP) guidelines. These approaches include obtaining consent through the care team to screen for VDD and consent for full trial enrollment simultaneously or separately, telephone consent, and deferred consent (for screening only). In participating sites where vitamin D status has or will be measured as the standard of care, or in patient sub-groups for whom vitamin D status is measured as part of standard of care, vitamin D status is determined clinically to confirm full trial eligibility. In participating sites and/or inpatient groups for whom vitamin D status is not measured as part of standard of care, vitamin D status is determined following deferred or informed consent (site-specific) using a research sample collected at time of screening—discard blood whenever possible. Vitamin D status is measured by site clinical labs, a research lab, or using an FDA- and Health Canada-approved point of care device (Qualigen® FastPack IP system). If 25OHD is less than 50 nmol/L, the patient is considered fully eligible and randomized into the trial.

Randomization

The Data Coordinating Centre (DCC) at the Ottawa Methods Centre at the Ottawa Hospital Research Institute (OHRI) has created a computer-generated randomization list. Participants are randomized 1:1 (intervention:control) using random variable block sizes to avoid major imbalances. Randomization is stratified by site to account for site-specific practice variation.

Allocation concealment and blinding

Randomization allocation is achieved using an electronic web page, where the participant is assigned an ID number and the site pharmacist receives an email with the ID number and treatment allocation. Randomization lists are only accessible to the Ottawa Methods Centre at the OHRI and research pharmacist(s). The active drug and placebo appear identical (e.g., color, consistency, volume, taste, smell, containers), in order to maintain blinding. Post-randomization, the research pharmacist dispenses the intervention (either vitamin D or placebo solution) to the nursing staff for administration. All trial personnel (the trial research coordinator, research assistants, site qualified investigators, principal investigator (PI), co-investigators, data management personnel, and statisticians), members of the health care team (treating physicians, bedside nurses, clinical pharmacists, etc.), and participants/families are blinded to the trial group assignment. The assigned intervention will not be revealed until all participants have been discharged from hospital, follow-up data collection and determination of research-related biochemical testing is complete, and the research database has been finalized. In the event of an emergency, blinding may be broken at the request of the clinical service after obtaining PI approval if unblinding would affect patient care. Blinding may also be broken at the request of the Data Safety Monitoring Board (DSMB) during interim analyses and at trial completion for the final analysis and manuscript preparation.

Intervention and controls

Participants randomized to the intervention arm receive an enteral cholecalciferol load (vitamin D3 (cholecalciferol) Oral Solution 50,000 IU/mL, Euro-Pharm Canada Inc.) at a dose of 10,000 IU/kg (maximum 400,000 IU).

Participants randomized to the control arm receive a placebo solution equivalent in volume to the appropriate dose of cholecalciferol at enrollment. The placebo is also provided by Europharm Canada International Inc.® All participants may also receive standard vitamin D dosing at the discretion of the care team (≤ 1000 IU/day), which is documented on the electronic case report form (eCRF). Baseline HRQL and Functional Status Scale (FSS) are obtained from parental interview within 72 h of trial enrollment. The hospital discharge (for participants with a hospital stay > 14 days, if not discharged within the 28 ± 7-day window), 28- and 90-day HRQL, and FSS measurements are obtained primarily in person, if participants remain hospitalized, or by telephone, email, or mail if participants have been discharged. Demographics, hospital course, adverse events, and health resource utilization are collected and documented on the eCRF.

Blood and urine sampling

Enrollment and up to two post-intervention (> 48 h, range: Days 2–7) urine and blood samples (through arterial or central venous lines or at the time of clinically indicated bloodwork) were collected from consecutively enrolled participants until July 2022. With consent from participating families, residual biosamples are maintained in the principal investigator’s laboratory at CHEO as part of the VITdALIZE-KIDS Trial Biobank. All policies and procedures surrounding this are consistent with the Organization for the Economic Co-Operation and Development (OECD) Guidelines on Human Biobanks and Genetics Research. The SPIRIT trial timeline is seen in Fig. 1.

Fig. 1

VITdALIZE-KIDS timeline of enrolment, interventions, and assessments

Outcomes

Our primary outcome is HRQL at 28 ± 7 days after intervention. This was identified in our survey of PICU families and caregivers as the most important outcome to them after survival [39] and is consistent with studies acknowledging HRQL as a preferred primary outcome for pediatric critical care trials [40]. HRQL is determined using the Pediatric Quality of Life Inventory Scale (PedsQL)™ 4.0 Generic Acute Scales [41] (ages 2–17) and the PedsQL™ Infant Scales (1–24 months). The PedsQL™ is one of the most widely used instruments for evaluating HRQL in pediatrics and has demonstrated reliability, sensitivity, validity, and responsiveness by self-reporting and/or proxy reporting in multiple pediatric populations including the PICU [42]. It has been used in numerous PICU studies [3, 43, 44] and was identified as one of the best HRQL tools for pediatric critical care research [2]. For infants under 1 month of age at enrollment, the baseline PedsQL™ questionnaires are not completed as no questionnaires are available or validated for that age. Proxy reporting is used for the primary outcome for consistency, but self-reporting is also completed when not prevented by age, ongoing critical illness, morbidity, or developmental status. Of note, as PICU patients often have comorbidities and impaired HRQL prior to PICU admission, pre-illness HRQL (prior to acute illness) is obtained from in person parental interview within 72 h of trial enrollment using the PedsQL™ (when baseline HRQL measurement is not feasible, bias in HRQL estimation is minimized when measurements are obtained at the earliest opportunity) [45]. Survival status is determined by review of the hospital chart and via follow-up calls; participants who cannot be contacted and have no in-hospital death recorded in their chart are considered lost to follow-up and survival is recorded as “unknown”. Participants who are deceased at follow-up are assigned a HRQL score of 0. Secondary and tertiary outcome measures are summarized in Table 2.

Table 2 Secondary and tertiary outcomes for the VITdALIZE-KIDS trial

Metabolomics sub-study

As short- and long-term clinical benefits are not mediated by cholecalciferol or 25OHD, but downstream changes in cell signaling and functioning of tissues/organs, we will perform essential translational research in a subset of 100 participants for whom enrollment and at least one set of post-intervention blood and urine samples have been collected > 48 h (range: Days 2–7, obtained as of July 2022). We will perform non-targeted metabolite profiling to better understand the effects of rapid VDD correction on human metabolism beyond calcium homeostasis, including stress response, antioxidant production, immunomodulation, and mitochondrial function [46]. Additionally, targeted metabolomics will be employed to better characterize changes following the single high-dose bolus by measuring vitamin D species beyond 25OHD [47] while also identifying new biomarkers that may predict inadequate vitamin D status or suboptimal responses to treatment [48].

Safety procedures

Participants are monitored for adverse events (AEs) from the time of intervention on a weekly basis during hospital admission (censored at 90 days) and through parent reporting at 28- and 90-day follow-ups. As the VITdALIZE-KIDS population is expected to experience AEs related to critical illness or its complications, a trial-specific definition of AEs is being used, as published by the CCCTG [49]. Specifically, in this trial, an AE is defined as any untoward medical occurrence that (1) is specified a priori as a safety endpoint (see below) or (2) is different from what is expected in the clinical course of the participant given their critical illness, co-morbidities, and related complications, AND may be related to the VITdALIZE-KIDS trial procedures (vitamin D administration and/or research procedures). Specified safety endpoints consist of the following AEs plausibly associated with vitamin D administration: (1) gastrointestinal bleeding (requiring transfusion) and/or perforation (requiring surgery) within 48 h of drug administration, (2) hypercalcemia persisting for more than and not occurring within 24 h of any parenteral bolus calcium administration AND one or more of renal failure requiring dialysis, nephrocalcinosis, hemodynamically significant arrhythmia, cardiorespiratory arrest, or death; and (3) nephrolithiasis with gross hematuria and/or nephrocalcinosis with renal failure. Should a trial participant develop significant symptoms related to vitamin D toxicity (potential drug-related adverse event), they are evaluated and followed by endocrinology and/or nephrology.

We follow Health Canada’s guidelines for reporting AEs, which entails reporting AEs if they are serious, related to the intervention, and unexpected given the participants’ critical illness, co-morbidities, related complications, and the known side effects of vitamin D toxicity. Similarly, we report AEs to local ethics boards according to local guidelines for reporting.

Sample size

The proposed sample size is 766 participants. Using the traditional t-test approach, 650 participants would provide 80% power (α = 0.05) to detect a 4.4 point difference in the PedsQL™ 4.0 [41] (SD = 20) between study arms at day 28, corresponding to the minimal clinically important difference (MCID). MCID is defined as the smallest difference in a score that patients perceive to be beneficial and that would ideally mandate a change in patient management; it was calculated in previous work (based on the standard error of measurement, SEM [50]) to be 4.5 points by parent proxy report and 4.4 points by child report [41]. Based on our pilot RCT data and related PICU literature [42, 44], we anticipate mean baseline (pre-illness) and 28-day PedsQL™ Total Scale Scores in this study population of approximately 75 and 50, respectively. Consequently, the proposed effect size of 4.4 points difference in 28-day (absolute) total score between intervention and control arms translates to a 20% improvement in HRQL. After discussion with our executive committee, we believed this effect size to be realistic given the strong associations documented between VDD and clinical outcomes in observational studies combined with eligibility criteria (validated during our VITdAL-PICU pilot study) and ensures recruitment of a sick PICU cohort that could benefit from vitamin D optimization.

Based on our pilot VITdAL-PICU study [34] and related PICU studies [43], we anticipated no more than a 15% loss to follow-up at 28 days resulting in ~ 650 evaluable trial participants from the 766. Note that while anticipated mortality is 5–10% (7/67 in the VITdAL-PICU pilot trial), participants for whom this death has been recorded in their hospital chart and/or is communicated at follow-up will not be considered lost to follow-up but will instead be assigned a HRQL score of 0.

Recruitment, compliance, and follow-up

We did not anticipate problems with participant compliance with the trial protocol. Participants are hospitalized and followed by the research staff, and the intervention is administered by nursing staff and documented in the participant’s medical record. Although we have planned for a liberal loss-to-follow up rate (15%) in our sample size calculation as participants are followed for 90 days after intervention, extensive efforts are put into participant retention (e.g., $25 gift cards for follow-up completions, collection of multiple points of contact for guardians and reminders sent via the preferred method). As ≤ 1000 IU/day vitamin D administration at care team discretion is allowable within our protocol (and an exclusion criterion is the intent of treating physician to administer doses above this for known clinical indications), we did not anticipate problems with physician compliance.

Data collection and management

Data collection is performed by research staff at each site, with 28- and 90-day follow-up assessments performed by the coordinating center staff for discharged participants. Data from questionnaires and medical records are entered into the web-based eCRF managed by The Ottawa Methods Centre—Electronic Data Capture System (OMC-EDCS) Platform, a validated EDCS platform used by over 100 studies in 15 countries for both regulated and non-regulated trials.

The following information is collected: demographic information and admission data, illness severity, relevant laboratory results and infection data, progression and resolution of organ dysfunction, fluid and inotrope administration, renal function and instances of calcium imbalance, occurrence of adverse events, length of mechanical ventilation, length of PICU and hospital stay, survival status, quality of life scores, and protocol deviations. All participant-related information (eCRFs, laboratory specimens, etc.) are kept strictly confidential. All records are kept in a secure, locked location and only research staff access the records. Participants are identified only by means of a coded number specific to each participant. The OMC-EDCS trial database is securely protected and encrypted. All computerized databases identify participants by numeric codes only.

Statistical analysis

Analyses will be performed using SAS® software (Cary, NC, USA) and will follow the intention-to-treat principle. A two-tailed p-value < 0.05 will be considered statistically significant. Study arms will be described and compared using (i) means with standard deviations or medians with inter-quartile range values for continuous variables or (ii) frequencies with percentages for categorical variables.

Primary outcome

The primary analytical plan will use the rank-based Mann–Whitney U test to compare 28-day PedsQL™ scores between the two arms. A non-parametric test is necessary as we do not expect PedQL™ data to be normally distributed, largely due to non-survivors being assigned a score of 0. Pre-specified secondary analysis of the primary outcome will also be done incorporating the PedsQL™ scores at baseline, 28 and 90 days. Participants who are lost to follow-up and do not complete the PedsQL™ at 28 days will be treated as missing in the analysis. Analysis will be done using a generalized linear repeated measures model accounting for repeated measurements on the same participant (a participant may die or be lost to follow-up). In this analysis, we will evaluate the relationship between the primary outcome measure and covariates such as baseline HRQL, 25OHD status, PedsQL™ age brackets, sex, kidney dysfunction according to RIFLE criteria [51], and illness severity measured using the pediatric risk of mortality (PRISM IV) score at enrollment. If potential relationships are observed, we will adjust for them in the analysis.

Secondary and tertiary outcomes

Other analyses will be evaluated between groups based on data type. Continuous outcome measures will be evaluated using the t-test or Mann–Whitney (where appropriate). Binary secondary outcome measures will be compared between the two treatment groups using Fisher’s exact or Chi-square test. For the analysis of outcomes measures that represent time to event (PICU or hospital discharge), we will apply the log rank test and generate Kaplan–Meier curves. The analysis may be expanded to multivariate regression models (e.g., if uneven distribution of important variables is identified). Participants who are lost to follow-up and do not complete the PedsQL™ at 90 days will be treated as missing in the analysis of this tertiary outcome.

Metabolomics analysis

Non-targeted metabolomics analysis will be performed on blood and urine using an extensively validated multiplexed separation method for high throughput metabolic phenotyping based on multisegment injection-capillary electrophoresis-mass spectrometry (MSI-CE-MS) [52, 53]. This method is applicable to the analysis of a wide array of ionic/polar metabolites in volume-limited biological samples with stringent quality control protocol [52], including matching serum and urine specimens. These will be analyzed for speciation of multiple vitamin D metabolites using LC–MS/MS with parallel reaction monitoring, including the major catabolites of 25OHD3 such as the 3-epimer of 25OHD3 and 24,25(OH)₂D3 [54]. This technique will also allow measurement of less bioactive vitamin D catabolites that are not typically resolved by conventional immunoassays [55] but which may be more impacted by high-dose bolus intervention, when compared with low-dose daily restoration regimens [56]. Furthermore, comprehensive lipidomic analyses of serum extracts will be performed using reversed-phase liquid chromatography-mass spectrometry (LC–MS) since increases in 25OHD have been shown to generate favorable changes in several distinct lipid classes coinciding with lower 28-day mortality, including sphingomyelins, plasmalogens, lysoplasmalogens, and lysophospholipids [57]. Additional previously unexplored metabolites of potential clinical significance will be identified by acquiring high resolution MS/MS with confirmation by comparison with authentic standards, thus allowing for their biochemical interpretation [58]. Complementary multivariate and univariate statistical methods will then be used to reduce data dimensionality while ranking metabolites significantly associated with changes in serum 25OHD concentrations and HRQL after adjustment for multi-hypothesis testing (false discovery rate; q < 0.05) and confounders. We will also report on and compare (by interventional group) the enrollment and post-intervention levels of calcium:creatinine.

Subgroup analyses

As with the primary outcome, we will perform additional exploratory subgroup analysis based on the following baseline subgroup characteristics: 25OHD status (grouped by screening levels 0–29 nmol/L and 30–49 nmol/L), age, sex, kidney dysfunction (according to RIFLE criteria), and illness severity (PRISM IV score).

Interim analyses and DSMB

Interim analyses are scheduled to occur after primary outcome data becomes available for 100 participants (completed), and 383 participants (50% of sample size) are enrolled and have completed all study procedures, with CRFs locked (not yet completed). The second interim analysis will also include a planned futility analysis. As per the Data Safety Monitoring Board (DSMB) Charter, the trial will be stopped for futility if the conditional power of the observed difference at the interim analysis is < 20%.

The DSMB is comprised of two clinical experts (pediatric intensivist and nephrologist) and a biostatistician (DSMB chair). Terms of Reference were created and approved by all DSMB members and the PI before recruitment commenced. The DSMB functions independently and at arm’s length from the trial investigators and the Steering Committee. The DSMB reviews trial deaths and adverse events potentially related to the trial intervention, makes recommendations to the Steering Committee based on these reviews regarding the continuation, modification, or termination of the trial, and comments on the relevance of new external published data from other trials that may impact on participant safety or efficacy of the trial treatments.

Monitoring

Sites are monitored once the first participant has been enrolled and primary outcome data obtained, then every 6 months (or more frequently as needed). Monitors ensure site adherence to protocol, GCP, and regulatory requirements.

Ethics and regulatory requirements

Approval from Health Canada and research ethics boards at all participating sites was obtained before the initiation of recruitment. Protocol amendments are communicated as necessary to those involved in the trial. Trial protocol was approved by Clinical Trials Ontario (CTO) (1761) and regulatory approval was obtained from Health Canada (control number 223325). The trial was first posted on Clinicaltrials.gov on November 15, 2018 (NCT03742505).

Close out

At trial close out, all research records will be retained for a minimum of 15 years. The trial drug will be returned to the coordinating center where it will be reconciled and destroyed according to site-specific pharmacy procedures and regulatory requirements. Trial data will only be accessible to the PI, research coordinator(s), or delegate.

Dissemination and access to data

The PI will submit the results of this trial for publication in a peer-reviewed journal. Metabolomics sub-study findings will be reported separately. Interim and final results will be presented at national and international scientific meetings, including at CCCTG meetings.

Data analysis will occur at the CHEO Research Institute (RI)/Ottawa Hospital Research Institute (OHRI).

Discussion

The prevalence and impact of VDD in critical illness is well established, both in Canada and around the world. Our group’s international phase II dose evaluation trial, VITdAL-PICU [34], confirmed that a single loading dose of 10,000 IU/kg of cholecalciferol (maximum 400,000 IU) can rapidly restore blood 25OHD levels in deficient patients to physiological levels, with no evidence of toxicity or related AEs. The VITdALIZE-KIDS trial is the first phase III, multicenter RCT investigating the hypothesis that rapid normalization of vitamin D status could represent a simple, inexpensive, and safe means of improving outcomes following pediatric critical illness. This trial is regulated by Health Canada and was confirmed to be compliant during a routine inspection by Health Canada (consisting of a sponsor inspection in May 2023 and two site inspections in June and September 2023).

To date, the DSMB has reviewed the first 100 participants enrolled with primary outcome data and did not have any concerns about safety, nor did they recommend any changes to the trial. However, during the recruitment phase of this trial, challenges to enrollment were encountered as outlined below. In light of these challenges, the Steering Committee and DSMB recommended that the trial continue at least until the interim analysis at 383 participants (50% enrollment target).

We had initially planned for 3 years of recruitment based on an expected 2–2.5 participants recruited per month per center (consistent with our experience across a wide variety of pediatric and adult RCTs) and assuming a VDD rate of 60–70%. However, we observed a lower-than-expected VDD rate of 41% (180 VDD out of 438 25OHD measured at the 100-participant interim analysis), resulting in an expected recruitment of an average of 1–2 participants per month per center, based on the current observed VDD rate. Additionally, due to the impact of COVID-19 related restrictions on family presence at the bedside and study staff presence and recruitment practice, it was necessary to make several modifications which have been documented by applying the CONSERVE protocol [59] and CONSERVE-SPIRIT checklist (Additional file 2). In light of these challenges, the recruitment duration has been amended to 6 years with a funding end date of March 2026.

Since initiating recruitment, we have also implemented several modifications to the protocol unrelated to COVID-19 that were aimed at removing barriers to enrollment and improving retention: (1) we have updated the length of stay (LOS) inclusion criterion from “expected PICU LOS > 48 h at trial enrollment” to the inclusion criterion of “PICU admission” and exclusion criterion of “expected PICU discharge prior to trial drug administration and/or expected hospital LOS < 3 days after enrollment.” This 3-day cutoff was determined after 2.5 years of recruitment (January 2022), at which point 4595 patients had been excluded for this criterion, but on review of hospital LOS of enrolled participants to that point, fewer than 5% of patients were actually discharged prior to 48–72 h. This amendment ensured that only critically ill patients would be enrolled, but the number of eligible patients could be maximized. (2) The inclusion criterion of post-intervention bloodwork at > 48 h (range: Days 2–7) was initially included to ensure sample collection for the metabolomics sub-study and interim DSMB analysis, but this was removed when sufficient patients had been enrolled to complete these (July 2022). Similarly, all post-intervention blood and urine collection also concluded at this point, to lessen the burden for families and trial sites. Additionally, more minor changes to the protocol are presented in Additional file 3.

Conclusion

High VDD rates in Canadian PICUs are associated with worse clinical outcome for patients and higher healthcare costs. Recognized interactions between vitamin D status and the health of multiple organ systems suggest this could represent an inexpensive, safe means of improving outcomes both in terms of patient health and PICU resource use. However, the effect of rapid normalization of vitamin D status on clinical outcomes following pediatric critical illness has never been evaluated in a clinical trial, prior to this RCT. Trial findings will be used to inform guidelines for vitamin D supplementation.

Trial status

Recruitment for the VITdALIZE-KIDS trial is expected to continue until March 2026 and was started at sites on the following dates: Children’s Hospital of Eastern Ontario (June 17, 2019), London Health Sciences Centre (September 26, 2019), Montreal Children’s Hospital (September 9, 2020), Centre hospitalier universitaire Sainte-Justine (November 18, 2020), McMaster Children’s Hospital (January 13, 2021), Alberta Children’s Hospital (February 10, 2021), BC Children’s Hospital (April 5, 2021), The Hospital for Sick Kids (December 13, 2021), Stollery Children’s Hospital (January 30, 2023), and IWK Health Centre (May 8, 2023). At the time of this manuscript development, recruitment has not yet started at Centre mère-enfant Soleil du CHU de Québec Université Laval. As of April 20, 2024, approximately 55% (n = 419) of the target sample size has been enrolled with approximately 98% (n = 411) of those having completed all study procedures, and the interim analysis has been scheduled. Protocol v25-Oct-2023 is currently in use.