Publication

• Title: Effect of High-Dose Vitamin D3 on Hospital Length of Stay in Critically Ill Patients With Vitamin D Deficiency: The VITdAL-ICU Randomized Clinical Trial
• Acronym: VITdAL-ICU
• Year: 2014
• Journal published in: JAMA
• Citation: Amrein K, Schnedl C, Holl A, Riedl R, Christopher KB, Pachler C, et al. Effect of high-dose vitamin D3 on hospital length of stay in critically ill patients with vitamin D deficiency: the VITdAL-ICU randomized clinical trial. JAMA. 2014;312(15):1520-1530.

Context & Rationale

• Background

  • Vitamin D deficiency is common in critically ill adults and may reflect pre-existing deficiency and critical illness-related changes (haemodilution, reduced binding proteins, altered metabolism, reduced enteral absorption).
  • Observational ICU cohorts link low 25-hydroxyvitamin D (25[OH]D) concentrations with higher mortality, infections and prolonged length of stay, but confounding and reverse causality limit causal inference.
  • Prior interventional evidence was limited and heterogeneous, leaving uncertainty about optimal dose, timing and whether any benefit is confined to severe deficiency.
• Research Question/Hypothesis

  • In adult ICU patients with vitamin D deficiency (25[OH]D ≤20 ng/mL), does rapid correction using high-dose cholecalciferol (vitamin D3) reduce hospital length of stay (primary endpoint) and improve mortality and other clinical outcomes (secondary endpoints) compared with placebo?
• Why This Matters

  • Vitamin D replacement is inexpensive, scalable and biologically plausible (immune and musculoskeletal effects); even modest outcome improvements would have major health-system implications.
  • High-dose bolus strategies can normalise levels quickly, but efficacy and safety in critical illness required randomised evidence.

Design & Methods

• Research Question: In critically ill adults with 25(OH)D ≤20 ng/mL, does high-dose cholecalciferol reduce hospital length of stay compared with placebo?
• Study Type: Randomised, double-blind, placebo-controlled, single-centre, stratified, parallel-group trial in 5 adult ICUs (Medical University of Graz, Austria); enrolment May 2010–March 2012.
• Population:
  • Setting: Medical, mixed surgical, neurological and cardiothoracic ICUs at a tertiary university hospital.
  • Key inclusion criteria: Adult ICU patients; 25(OH)D ≤20 ng/mL at study inclusion; anticipated ICU stay ≥48 hours; enteral access/ability to receive study drug.
  • Key exclusion criteria: Hypercalcaemia at screening (total calcium >10.6 mg/dL or ionised calcium >5.4 mg/dL); nephrolithiasis; granulomatous disease/tuberculosis/sarcoidosis; pregnancy/lactation; competing trial enrolment; inability to take study medication.
  • Baseline vitamin D status: Mean 25(OH)D 13.0±4.4 ng/mL (vitamin D3) vs 13.1±4.7 ng/mL (placebo); 42.1% had severe deficiency (≤12 ng/mL).
• Intervention:
  • Cholecalciferol (vitamin D3): 540,000 IU enterally as a loading dose, followed by 90,000 IU enterally once monthly for 5 months (maintenance), in addition to usual ICU care.
  • Permitted background intake: Standard vitamin D provided via enteral/parenteral nutrition (approximately 200 IU/day) allowed; additional vitamin D outside the protocol was discouraged.
• Comparison:
  • Placebo: Matching placebo (vehicle) administered enterally using the same loading and monthly maintenance schedule, plus usual ICU care.
• Blinding: Double-blind (participants, treating clinicians, investigators and outcome assessors); study drug and placebo matched; allocation code held by pharmacy.
• Statistics: Sample size 468 required to detect a 2-day reduction in mean hospital length of stay (two-sided alpha 0.05; 80% power [beta 0.20]); inflated to 490 to allow ~5% dropout; primary analyses performed in a modified intention-to-treat population (final analyses n=475); primary endpoint compared with Mann–Whitney U test; mortality analysed with Cox proportional hazards models.
• Follow-Up Period: In-hospital outcomes through discharge/death; mortality assessed at ICU discharge, 28 days, hospital discharge and 6 months; biochemical follow-up at days 3, 7 and 28 and at 6 months; functional outcomes assessed at 6 months in survivors where feasible.

Key Results

This trial was not stopped early. 492 patients were randomised; 475 were included in the final analyses (vitamin D3 n=237; placebo n=238).

Outcome	Vitamin D3	Placebo	Effect	p value / 95% CI	Notes
Hospital length of stay, days (median [range])	19.3 (0–693)	20.1 (0–679)	Not reported	P=0.98	Primary endpoint; Mann–Whitney U test
ICU length of stay, days (median [range])	10.7 (0–693)	9.6 (0–679)	Not reported	P=0.38	Secondary endpoint; Mann–Whitney U test
ICU mortality	54/237 (22.8%)	63/238 (26.5%)	HR 0.97	95% CI 0.67 to 1.39; P=0.86	Cox proportional hazards model
28-day mortality	52/237 (21.9%)	68/238 (28.6%)	HR 0.76	95% CI 0.53 to 1.09; P=0.14	Secondary endpoint; Cox proportional hazards model
Hospital mortality	67/237 (28.3%)	84/238 (35.3%)	HR 0.81	95% CI 0.58 to 1.11; P=0.18	Secondary endpoint; Cox proportional hazards model
6-month mortality	83/237 (35.0%)	102/238 (42.9%)	HR 0.78	95% CI 0.58 to 1.04; P=0.09	Secondary endpoint; Cox proportional hazards model
Serum 25(OH)D at day 7, ng/mL (mean ± SD)	35.5 ± 20.6	14.5 ± 5.1	Not reported	P<0.001	Biochemical separation (day 7 samples: 202 per group)
Total serum calcium >10.6 mg/dL at 6 months	4/37 (11%)	1/43 (2%)	Not reported	Not reported	Safety; measured in subset at 6 months
Falls by 6 months	33/136 (24.3%)	27/153 (17.7%)	Not reported	P=0.17	6-month follow-up subset (reported in supplement)
Fractures by 6 months	2/136 (1.5%)	2/153 (1.3%)	Not reported	P=0.86	6-month follow-up subset (reported in supplement)

• Primary endpoint neutral: High-dose vitamin D3 did not reduce hospital length of stay (19.3 vs 20.1 days; P=0.98) or ICU length of stay (10.7 vs 9.6 days; P=0.38).
• Mortality signals were secondary and mostly imprecise: Mortality was numerically lower with vitamin D3 but not statistically significant overall (6-month mortality HR 0.78; 95% CI 0.58 to 1.04; P=0.09); in severe deficiency (≤12 ng/mL), vitamin D3 was associated with lower hospital mortality (28.6% vs 46.1%; HR 0.56; 95% CI 0.35 to 0.90; P=0.01; interaction P=0.04) and lower 28-day mortality (20.4% vs 36.3%; HR 0.52; 95% CI 0.30 to 0.89; P=0.02).
• Biochemical separation achieved with limited safety signal: Day 7 25(OH)D rose to 35.5±20.6 ng/mL vs 14.5±5.1 ng/mL (P<0.001) and 52.0% exceeded 30 ng/mL by day 7 (vs 1.5% placebo); hypercalcaemia at 6 months occurred in 11% vs 2% among those tested, and falls/fractures were not significantly different.

Internal Validity

• Randomisation and allocation concealment: Computer-generated sequence with block size 8, stratified by ICU type and sex; pharmacy-controlled packaging suggests adequate concealment and reduced selection bias.
• Dropout/exclusions and analysis set: 492 randomised; 480 received study drug (240 per arm); 5 withdrew consent and 12 did not receive study medication (placebo 3; vitamin D3 9); final analyses included 475 (placebo 238; vitamin D3 237), indicating a modified intention-to-treat approach rather than strict ITT.
• Performance/detection bias: Double blinding with matched preparations reduces performance bias; mortality outcomes are objective; hospital length of stay is objective but influenced by organisational discharge processes and competing risk of death.
• Protocol adherence and separation of the variable of interest: Biochemical separation was substantial: day 3 25(OH)D 33.5±17.6 vs 13.9±4.9 ng/mL; day 7 35.5±20.6 vs 14.5±5.1; day 28 32.7±15.7 vs 17.3±8.7; month 6 46.0±20.3 vs 26.2±17.8; however only 52% exceeded 30 ng/mL by day 7, reflecting variability and incomplete normalisation in many patients.
• Baseline characteristics: Groups were well balanced (age 63.9±15.5 vs 65.3±14.0 years; SAPS II 32.4±14.3 vs 34.2±14.7; mechanical ventilation at inclusion 63.7% vs 64.7%; norepinephrine use 55.3% vs 52.9%).
• Heterogeneity and timing: Broad ICU case-mix increases biological heterogeneity; randomisation occurred a median 2.1 days after ICU admission (mean 3.0 days), which may be late for mechanisms requiring very early immunomodulation.
• Outcome missingness: Long-term biochemical and functional outcomes were available in subsets (e.g., month 6 laboratory data in 43 placebo vs 37 vitamin D3), limiting precision for safety and longer-term endpoints.
• Statistical rigour: Planned sample size 490; analysed 475 (slightly reduced power for primary endpoint); mortality outcomes were secondary; subgroup findings (severe deficiency) included interaction testing but remain vulnerable to multiplicity.

Conclusion on Internal Validity: Overall, internal validity appears moderate-to-strong given robust stratified randomisation, allocation concealment, blinding and clear biochemical separation, but is limited by a modified ITT population, incomplete long-term follow-up data and an organisationally sensitive primary endpoint (hospital length of stay).

External Validity

• Population representativeness: Adult, single-centre European ICU cohort (100% White) with documented deficiency (≤20 ng/mL) and anticipated ICU stay ≥48 hours; generalisability may be limited for more ethnically diverse populations, shorter-stay ICU cohorts and systems with different discharge practices.
• Applicability: Enteral high-dose cholecalciferol is inexpensive and operationally feasible; however, routine ICU screening for 25(OH)D and rapid treatment pathways vary by health system, and effects may differ by baseline deficiency severity and timing of administration.

Conclusion on External Validity: Findings are reasonably applicable to similar tertiary adult ICUs caring for vitamin D–deficient patients, but external validity is constrained by single-centre design, homogeneous ethnicity and the dependence of length-of-stay outcomes on local practice.

Strengths & Limitations

• Strengths:
  • Randomised, double-blind, placebo-controlled design with pharmacy-controlled allocation.
  • Large ICU supplementation trial for its era, with stratification by ICU type and sex, reducing risk of imbalance across heterogeneous ICUs.
  • Robust biological verification of exposure (serial 25[OH]D and mineral metabolism measurements), demonstrating clear separation between groups.
  • Objective mortality outcomes with follow-up to 6 months; additional survivor-centred functional outcomes collected where feasible.
• Limitations:
  • Single-centre design and 100% White cohort limit generalisability.
  • Primary endpoint (hospital length of stay) is sensitive to discharge processes and competing risk of death.
  • Modified ITT analysis set (475/492 randomised) introduces potential bias if exclusions were prognostically relevant.
  • Underpowered for mortality; statistically significant mortality effects were confined to a severe deficiency subgroup and require confirmation.
  • Long-term laboratory and functional outcomes were available only in subsets, limiting precision for safety and longer-term endpoints.

Interpretation & Why It Matters

• Clinical implications

  • Routine high-dose cholecalciferol to shorten hospital stay in vitamin D–deficient ICU patients is not supported by the primary outcome.
  • The consistent biochemical correction confirms feasibility of rapid repletion, but clinical benefit (especially mortality) was not definitively established and was not the basis for powering.
  • The severe deficiency subgroup signal (≤12 ng/mL) is hypothesis-generating and suggests that future trials should enrich for profound deficiency and prioritise patient-centred endpoints (mortality, ventilator-free days) rather than length of stay alone.

Controversies & Other Evidence

• Primary endpoint choice: Hospital length of stay is multifactorial and influenced by discharge processes; a neutral LOS finding does not exclude biologically meaningful effects on mortality or organ dysfunction, but also makes the trial vulnerable to context-specific practice variation.
• Modified ITT and post-randomisation exclusions: Excluding 17/492 randomised patients (5 consent withdrawals; 12 not treated) can bias effect estimates if exclusions correlate with prognosis, particularly when mortality is an outcome.
• Severe deficiency subgroup: Mortality benefit was confined to 25(OH)D ≤12 ng/mL (hospital mortality interaction P=0.04); despite interaction testing, multiplicity and subgroup fragility limit certainty, and this subgroup effect is not a sufficient basis for practice change without confirmation.
• Subsequent large multicentre RCT (VIOLET): In vitamin D–deficient critically ill adults randomised to a single enteral 540,000 IU vitamin D3 dose vs placebo within 12 hours of ICU admission decision, 90-day mortality was 23.5% vs 20.6% (RR 1.17; 95% CI 0.96 to 1.43; P=0.26), and the trial was stopped early for futility, challenging the generalisability of subgroup mortality benefits seen in VITdAL-ICU.¹
• Updated systematic review/meta-analysis: An updated meta-analysis of randomised trials reported reduced overall mortality with vitamin D administration (RR 0.78; 95% CI 0.62 to 0.97; P=0.03), shorter ICU length of stay (mean difference −3.13 days; 95% CI −5.36 to −0.90; P=0.006) and shorter duration of mechanical ventilation (mean difference −5.07 days; 95% CI −9.68 to −0.47; P=0.03); heterogeneity across dosing strategies, routes, baseline deficiency definitions and trial quality remains a central interpretive challenge.²

Summary

• In adults with 25(OH)D ≤20 ng/mL in ICU, high-dose cholecalciferol (540,000 IU loading + monthly maintenance) rapidly increased 25(OH)D (day 7: 35.5±20.6 vs 14.5±5.1 ng/mL; P<0.001).
• The primary outcome was neutral: hospital length of stay was unchanged (median 19.3 vs 20.1 days; P=0.98).
• Mortality was numerically lower with vitamin D3 overall but not statistically significant; in severe deficiency (≤12 ng/mL), mortality reductions were observed (hospital mortality HR 0.56; 95% CI 0.35 to 0.90; P=0.01; interaction P=0.04).
• Safety signals were limited but incomplete: hypercalcaemia at 6 months occurred in 11% vs 2% among those tested; falls and fractures were not significantly different.
• Later multicentre evidence (VIOLET) did not demonstrate benefit and stopped early for futility; pooled evidence remains heterogeneous, reinforcing the need for targeted trials in profoundly deficient patients.

Overall Takeaway

VITdAL-ICU demonstrated that aggressive enteral vitamin D3 repletion can correct biochemical deficiency rapidly in critically ill adults, but it did not improve the primary endpoint of hospital length of stay. Mortality signals—especially in profound deficiency—were intriguing yet secondary and not definitive, and subsequent multicentre data have not confirmed a clinical benefit, so routine high-dose bolus vitamin D3 to improve ICU outcomes is not justified on this trial alone.

Bibliography

• 1.Ginde AA, Brower RG, Caterino JM, et al. High-dose vitamin D3 for critically ill vitamin D-deficient patients. N Engl J Med. 2019;381(26):2529-2540. Link
• 2.Menger JMM, Lee ZY, Notz Q, Wallqvist J, Hasan MS, Elke G, et al. Administration of vitamin D and its metabolites in critically ill adult patients: an updated systematic review with meta-analysis of randomized controlled trials. Crit Care. 2022;26:268. Link