Vitamin D and ICU delirium: strong associations, elusive proof of causation

Vitamin D deficiency roughly doubles the risk of delirium in hospitalized and ICU patients, according to converging observational, meta-analytic, and genetic evidence. A 2023 meta-analysis of nine studies found severe deficiency (<25 nmol/L) carried an odds ratio of 1.98 for delirium compared with normal levels, (pubMed central) (pubMed) while Mendelian randomization studies in over 300,000 UK Biobank participants suggest this relationship is causal (HR 0.74 per 10 nmol/L genetically instrumented increase, p = 0.0004). (PubMed) (Neurology) Yet a critical gap persists: no randomized controlled trial has tested vitamin D supplementation specifically for ICU delirium prevention. Meanwhile, 40-70% of ICU patients are vitamin D deficient at admission, (PubMed central) (ResearchGate) making this a potentially modifiable risk factor of enormous scale. The biological plausibility is strong — vitamin D modulates at least five of the six established pathogenic pathways for delirium (journal of cardiothoradc a...) — but the supplementation evidence needed to close the loop remains absent.

The epidemiological case is substantial but not unanimous

The observational literature tilts strongly toward an inverse association between vitamin D levels and delirium incidence, though a handful of smaller studies found null results. The most rigorous synthesis comes from Fu et al. (2023, PLOS ONE), who pooled nine observational studies and found a dose-dependent relationship: severe vitamin D deficiency (<25 nmol/L) versus normal (>75 nmol/L) yielded an OR of1.98 (95% Cl: 1.412.79; I² = 25%), (PubMed Central) while moderate deficiency (25-50 nmol/L) produced an OR of1.50 (95% Cl: 1.12-2.00). (PubMed central) Insufficiency (50-75 nmol/L) did not reach significance, (PubMed central) suggesting a threshold effect rather than a linear gradient at higher levels. (PubMed)

Individual studies reinforce this pattern with varying effect sizes. Velayati et al. (2020) followed 398 CABG patients and found severe deficiency (<10 ng/mL) tripled delirium odds (Journal of Cardiothoradc a.TT) (OR 3.18,95% Cl: 1.29-7.78) (Journal of Cardiothoradc a.T7) after multivariate adjustment. (scienceDirect) (Perfusion.com) Qiu et al. (2022), studying 632 surgical ICU patients at the Cleveland Clinic, identified 30 ng/mL as a critical threshold (PubMed) — below which delirium risk climbed linearly, above which no further protection was observed. (PubMed) Among 310 critically ill elderly COVID-19 patients, Gholi et al. (2022) found vitamin D deficiency carried a 54% higher delirium hazard (HR 1.54,95% Cl: 1.022.33). (PubMed) (scienceDirect) In 1,029 elderly hip fracture patients, deficiency increased delirium odds by 52% (OR 1.52,95% Cl: 1.01-2.31). (PubMed central)

Counterbalancing these positive findings, Morandi et al. (2013) in a small Vanderbilt cohort of 120 medical ICU patients found no association (pubMed) (OR 1.01,95% Cl: 0.981.02), (PubMed) and Jia et al. (2024) in a secondary analysis of390 cardiac surgery patients similarly reported null results (adjusted OR 0.99 per 1 ng/mL increase).(PubMed central) These negative studies were notably smaller and assessed delirium over shorter time windows. A separate meta-analysis by Hung et al. (2022, J Clin Anesth), pooling seven studies with 2,673 patients, confirmed that preoperative deficiency significantly raised the risk of postoperative delirium and cognitive dysfunction (scienceDirect) (pooled OR 1.54, p < 0.01). (ResearchGate)

Mendelian randomization provides the strongest causal signal

The gold-standard evidence for causality in this domain comes from two Mendelian randomization analyses exploiting the natural randomization of vitamin D-related genetic variants. Bowman et al. (2019, Neurology) analyzed 313,121 European-descent participants aged >60 from the UK Biobank, identifying 544 incident delirium cases. (Neurology) (Neurology) Genetic variants that raise circulating vitamin D were protective: HR 0.74 per 10 nmol/L increase (95% Cl: 0.62-0.87; p = 0.0004), with no evidence of pleiotropy confounding the result. (PubMed) (Neurology)This design minimizes reverse causation — the possibility that illness causes low vitamin D rather than the reverse.

(PubMed Central)

Pilling et al. (2021, JAGS) extended this work with 351,320 participants and up to 14 years of follow-up, finding 3,634 incident delirium cases. Observationally, vitamin D deficiency (<25 nmol/L) versus sufficiency (>50 nmol/L) carried a hazard ratio of2.49 (95% Cl: 2.242.76; p = 3 x 10'⁶⁸). (PubMed Central) (wiiey Online Library) Instrumental variable analysis again confirmed the genetic association. Together, these studies represent the most compelling evidence that low vitamin D causally contributes to delirium risk, though they reflect community-dwelling populations rather than ICU cohorts specifically.

Five convergent biological mechanisms explain why vitamin D matters for the brain

The pathophysiology of delirium involves neuroinflammation, cholinergic deficiency, blood-brain barrier disruption, oxidative stress, and neurotransmitter imbalance.

(PubMed central) Vitamin D directly modulates at least five of these pathways, providing robust biological plausibility.

Neuroinflammation suppression is perhaps the most potent mechanism. The vitamin D receptor (VDR) physically binds IKKp protein, blocking TNF-a-induced NF-kB activation (pubMed central) — the master inflammatory transcription factor. Vitamin D downregulates pro-inflammatory cytokines IL-6, TNF-a, and IL-ip while upregulating anti-inflammatory IL-10. (sdenceDirect) (sdenceDirect) It suppresses the NLRP3 inflammasome and shifts microglia from pro-inflammatory Ml to neuroprotective M2 phenotype (PubMed central) via TLR4/MyD88/NF-KB pathway inhibition. Since the neuroinflammatory cascade — peripheral immune signals activating brain parenchymal cells — is considered a primary driver of ICU delirium, (sciencePirect) this mechanism is directly relevant, (plos +2)

Cholinergic enhancement addresses the leading neurotransmitter hypothesis of delirium. The active form of vitamin D (l,25(OH)₂D₃) elevates choline acetyltransferase (ChAT) activity (mppi) by 12-45% in specific brain nuclei, including the arcuate-median eminence and bed nucleus of the stria terminalis (PubMed) (Sonnenberg et al., 1986). VDR-knockout mice show significant decreases in brain acetylcholine levels. (mppi ) Vitamin D also modulates acetylcholinesterase activity and prevents pathological AChE increases in Alzheimer's models. (Taylor & Francis online) This is critical because cholinergic deficiency is the dominant neurotransmitter theory of delirium, (PubMed central) and anticholinergic medications are established delirium précipitants. (Frontiers) (sciencePirect)

Blood-brain barrier integrity is maintained by vitamin D through upregulation of tight junction proteins — ZO-1, occludin, and claudin-5. Won et al. (2015) demonstrated that calcitriol prevented hypoxia-induced barrier dysfunction in brain endothelial cells (plos) via VDR-mediated NF-kB signaling. (Frontiers) Vitamin D also regulates aquaporin-4, the principal brain water channel protein implicated in glymphatic clearance of neurotoxic waste, (journal of Cardiothoradc a... ) BBB disruption permits peripheral inflammatory mediators to access brain parenchyma, triggering the neuroinflammatory cascade central to delirium.

(Frontiers +3)

Neurotrophic support operates through vitamin D's upregulation of nerve growth factor (NGF), brain-derived neurotrophic factor (BDNF), and glial cell-derived neurotrophic factor (GDNF). (PubMed central) Direct intrahippocampal injection of calcitriol induces NGF expression in adult rats. (PubMed central) NGF is critical for survival of basal forebrain cholinergic neurons — the same neurons whose dysfunction underlies delirium. BDNF supplementation via vitamin D in aging rats mitigated age-related cognitive decline.

(PubMed)

Oxidative stress reduction rounds out the mechanistic picture. Vitamin D activates the Nrf2/HO-l antioxidant pathway, (PubMed centrai) (PubMed central) upregulates glutathione, superoxide dismutase, catalase, and glutathione peroxidase, and inhibits inducible nitric oxide synthase in microglia, (ncbi+2) ICU patients face overwhelming oxidative stress from sepsis, surgery, and mechanical ventilation. Vitamin D's antioxidant actions directly counter this pathogenic mechanism.

Vitamin D deficiency is near-universal in the ICU

The prevalence data underscore the scale of the problem. Across multiple studies and geographic regions, 40-70% of ICU patients have vitamin D deficiency (25(OH)D < 20 ng/mL), (PubMed central) (Nature) with typical median levels of 18-22 ng/mL. A Taiwanese multicenter study of538 patients reported a median of 18.3 ng/mL (IQR 13.7-23.9), with 59% deficient — compared to 22.4% in the general population of the same region.

(PubMed central) In critically ill cancer patients, deficiency prevalence reaches 74%, and severe deficiency (<12 ng/mL) affects 54%.(mdpi)

Septic patients fare worst: (Biomedcentrai) a Korean cohort reported median levels of just 13 ng/mL, ( mdpi) while Australian data showed sepsis patients at 7.9 ± 3 ng/mL versus 24.6 ± 6.7 ng/mL in controls. (ResearchGate) Contributing factors include pre-existing deficiency from chronic disease and limited sun exposure, acute metabolic dysrégulation during critical illness, hepatic and renal dysfunction impairing vitamin D hydroxylation, hemodilution from aggressive fluid resuscitation, and impaired gastrointestinal absorption. (PubMed central) Notably, vitamin D levels drop rapidly after ICU admission independent of pre-illness status, suggesting that critical illness itself depletes vitamin D stores.

(PubMed Central)

The VDR is expressed widely in the brain — in neurons and glia of the hippocampus (CA1, CA2), hypothalamus, substantia nigra, cortex, amygdala, and thalamus (PubMed) (nlh) — confirming that vitamin D functions as a neurosteroid with direct CNS activity.

(journal of Cardiothoracic a... ) The brain also possesses local activation machinery (CYP27B1 in microglia and pericytes), (nlh) meaning that peripheral deficiency likely translates to central deficiency, (mdpi)

No data exist on vitamin D and delirium subtypes

Despite the clear clinical relevance, no published study has examined whether vitamin D levels differentially predict hyperactive, hypoactive, or mixed delirium. All existing studies treat delirium as a binary outcome. This is a notable gap because the subtypes differ substantially in prevalence and prognosis. In a systematic review of 131 studies comprising 13,902 delirious ICU patients (Krewulak et al., 2022), hypoactive delirium was the most common subtype at 50.3%, followed by mixed (27.7%) and hyperactive (22.7%). Mixed delirium carried the worst outcomes across all metrics. (PubMed central) (springer) Given that hypoactive delirium is associated with older age and worse pre-morbid cognition (oxford Academic) — populations also at highest risk for vitamin D deficiency — a mechanistic link is plausible but undemonstrated.

Sepsis encephalopathy and post-ICU cognition remain underexplored

Sepsis-associated encephalopathy (SAE), affecting up to 50% of sepsis patients, (Nature) shares pathophysiological mechanisms with delirium including neuroinflammation, BBB disruption, and microglial activation. (PubMed) (PubMed central) Preclinical evidence is promising: Sun et al. (2025) showed that calcitriol reduced neuronal pyroptosis markers by -40% and increased survival from 41% to 72% in a mouse sepsis model, with corresponding cognitive improvement. (PubMed) (sdenceDirect) However, no human clinical trial has tested vitamin D for SAE prevention or treatment.

For post-ICU cognitive impairment, the most relevant evidence comes from the VIOLET-BUD trial (Han et al., 2021), an ancillary study of the large VIOLET RCT. Among 95 vitamin D-deficient critically ill adults randomized to 540,000 IU vitamin D₃ versus placebo, global cognition scores at approximately 14 months were nearly identical (79.6 vs. 82.1; adjusted OR 0.83,95% Cl: 0.50-1.38). (Eisevierpure) One executive function subscale was actually lower in the treatment group. (SdenceDirect) This null result dampens enthusiasm for highdose supplementation as a cognitive rescue strategy, though the small sample size, singledose design, and late outcome assessment limit its conclusions.

Vitamin D's relationship with sepsis outcomes more broadly is better established. De Haan et al. (2014) pooled 14 observational studies (n = 9,715) and found deficiency increased infection risk (RR 1.49), sepsis risk (RR 1.46), and in-hospital mortality (RR 1.79). (PubMed) (PubMed central) The VITdAL-ICU trial (Amrein et al., 2014) showed no overall mortality benefit from high-dose supplementation, (Nature)but a pre-specified subgroup with severe deficiency (<12 ng/mL) showed hospital mortality of 28.6% versus 46.1% (PubMed) (HR 0.56,95% Cl: 0.35-0.90) — a hypothesis-generating finding that has not been confirmed in larger trials. (PubMed Central)

Conclusion

The evidence linking vitamin D deficiency to ICU delirium is compelling across multiple lines of inquiry but incomplete at the interventional level. Observational data consistently show 1.5- to 3-fold increases in delirium risk among deficient patients, (PubMed central)

(plos) Mendelian randomization supports causality, (Neurology) and five distinct biological mechanisms provide robust plausibility. The near-universal prevalence of vitamin D deficiency in ICU populations (40-70%) (PubMed central) makes this a potentially highimpact modifiable risk factor. (Nature+2) Yet the field lacks the decisive evidence: a well-powered RCT testing vitamin D supplementation for ICU delirium prevention as a primary endpoint. The failure of the VIOLET-BUD ancillary study to improve post-ICU cognition — though limited in design — adds a note of caution about translating associative evidence into therapeutic benefit. (sdenceDirect) Future trials will need to address timing of supplementation (pre-admission vs. early ICU), dosing strategies (loading dose vs. daily), baseline vitamin D stratification, and delirium subtype differentiation. (Frontiers) Until such trials are completed, vitamin D deficiency should be viewed as a robust risk marker for ICU delirium, with correction of severe deficiency a reasonable clinical practice given its low cost and safety profile, (mdpi) even as definitive proof of delirium prevention remains

Outstanding. (PubMed Central) (Neurology)