Association between vitamin D concentration and delirium in hospitalized patients: A meta-analysis.

Delirium, occurs in hospitalized patients more often than in the elderly. It mainly shows disorders of consciousness, disorganized thinking, lack of attention, emotional disorders, etc., and is often accompanied by other symptoms, and severe cases are even life-threatening [ 1 ]. The pathophysiological causes of delirium remain unclear but are generally thought to be related to acute infections, drugs, general anesthesia, surgery, endocrine disorders, cardiovascular disease, and several other factors (hypoxic damage to the brain, poisoning) [ 2 ]. Available data suggest that deficiencies of omega-3 polyunsaturated fatty acids and B vitamins cause cardiovascular and neurological disorders, including paresthesias, memory loss, and dementia, among others [ 3 , 4 ]. Omega-3 polyunsaturated fatty acids, like vitamins and minerals, are essential for the human body and have anti-inflammatory and vasodilatory properties, which may protect blood vessels and maintain blood perfusion in the brain. Meanwhile, there is increasing evidence suggesting that vitamin D deficiency may have an association with delirium [ 5 , 6 ]. Vitamin D is a fat-soluble vitamin that has many physiological functions in the human body, such as promoting bone growth, regulating calcium and phosphorus metabolism, regulating immune function, and so on [ 7 ]. Several studies have also confirmed the potential role of vitamin D in neurological and neuropsychiatric disorders, and vitamin D may be an important regulator of brain development [ 8 , 9 ]. In addition to this, there is increasing evidence that the vitamin D metabolic pathway may play an important role in obstetric and gynaecological diseases and reproduction, such as endometriosis, ovarian cancer, polycystic ovary syndrome (PCOS), and even breast cancer [ 10 , 11 ]. There are currently no specific guidelines for vitamin D supplementation in affected women, but there should be selectivity in the decision to supplement vitamin D, as its excess may have deleterious effects on fertility [ 10 ]. 25 hydroxyvitamin D (25 (OH) D) is the intermediate form of vitamin D that is converted in the liver, and it is generally considered as a marker to assess whether vitamin D is deficient [ 12 ]. According to the American Endocrine Society guidelines, a serum 25 (OH) D concentration of 25 nmol/L was defined as severe deficiency, a serum 25 (OH) D concentration of 25–50 nmol/L as deficiency, a serum 25 (OH) D concentration of 50–75 nmol/Las insufficiency, and a serum 25 (OH) D concentration of > 75 nmol/L as normal [ 13 , 14 ]. Its role in calcium homeostasis and bone has been extensively studied. Of course, there are also some studies showing some effects of vitamin D on cognition and delirium [ 5 , 15 , 16 ]. A review by Groves et al summarizes the association of vitamin D deficiency in adulthood with brain related adverse outcomes such as neuropsychiatric disorders and neurodegenerative diseases [ 17 ]. Qiu et al reported that preoperative vitamin D deficiency was closely related to postoperative delirium [ 14 ]. Ford et al. found that vitamin D levels are lower in patients with delirium compared to those without delirium [ 18 ]. Pilling et al. support that low levels of vitamin D are a risk factor for the delirium [ 19 ]. However, Morandi et al. concluded that low concentrations of vitamin D levels were not correlated with whether delirium developed and that 25 (OH) D levels in patients with delirium were not significantly different from those in healthy patients [ 20 ]. Several studies also did not find an association between low levels of vitamin D and cognitive decline [ 15 , 21 , 22 ]. To date, whether serum vitamin D deficiency is a risk factor for delirium has not reached a consensus, and no meta-analysis of these issues has been performed. A meta-analysis is urgently needed to resolve these controversies. Thus, we conducted a meta-analysis to investigate the association between vitamin D concentration and inpatient delirium. The primary outcome of this meta-analysis was the association between low levels of vitamin D and the risk of delirium in the inpatient setting, and the other outcomes were the relationships between other potential risk factors and the risk of delirium. We hypothesized that low vitamin D concentrations are a risk factor for delirium in hospitalized patients.

According to the comprehensive search of the above databases, a total of 314 pieces of literature were screened. After checking to remove duplicates, 278 articles remained. Then, based on the inclusion and exclusion criteria, filter the title and abstract, and finally, select 12 articles. 3 articles [ 6 , 18 , 24 ] were excluded after a full-text search, 9 observational studies did the quantitative analysis. One [ 25 ] was excluded because it only provided relevant OR values, the outcome measures did not correspond, and attempts to contact the authors did not obtain the data. Finally, 8 pieces of literature [ 14 , 19 , 20 , 26 – 30 ] were included in the quantitative synthesis in the meta-analysis ( Fig 1 ) . For the final quantitative analysis, eight observational studies (three prospective, four retrospective, and one case-control study) were included. The years of publication of these observational studies were between 2013 and 2021. In one [ 20 ] of them, there is no detailed data in the literature, so only OR values were extracted. The NOS scale was used to assess the quality of non-RCT studies, with a total of eight entries and nine assigned points (from 0 to 9 stars). The basic characteristics and NOS scores of the included articles are presented in Tables 1 and 2 . After repeated evaluation by multiple researchers, the average score of NOS is 7.8, which can be considered as high-quality literature. Abbreviations CAM-ICU: Confusion Assessment Method for the ICU NA: not available. A high-quality selection was considered a star. The more stars assigned to a study, the better the quality of the study. The "comparability" category receives a maximum of two stars. In the “selection” and "outcome/ exposure" categories, each numbered item can receive a maximum of one star. No publication bias was done in this paper due to the small number of included works of literature. When heterogeneity was high, a sensitivity analysis was performed. After excluding the literature by article, the results were consistent before and after the outcome measures, which indicated that the results were robust and credible ( Fig 2 ). As shown in Fig 3 , the incidence of delirium was higher in patients with severe deficiency of serum vitamin D than in those with normal vitamin D levels (OR: 1.98; 95% CI: 1.41 to 2.79; I 2 = 25%, P<0.001). When serum vitamin D was deficient, the incidence of delirium was also greatly different between cases and controls (OR: 1.50; 95% CI: 1.12 to 2.00; I 2 = 0.0%, P = 0.006). However, when serum vitamin D insufficiency (50–75 nmol/L) was compared with normal serum vitamin D (> 75 nmol/L), there was not statistical difference between the two groups (OR: 1.29; 95% CI: 0.97 to 1.72; I 2 = 0.0%, P = 0.08). Heterogeneity I 2 of less than 50% was observed at the time of meta-analysis, so we adopted a fixed-effects model. A: <25nmol/L. B: 25-50nmol/L. C: 50-75nmol/L. As can be seen in Fig 4 , Five studies compared the incidence of delirium using a cut-off value of 25 nmol/L and showed a marked difference between groups for serum vitamin D 25 nmol/L (OR: 1.52; 95%CI: 1.40–1.64; I 2 = 0%, P<0.001). Six observational studies reported the incidence of delirium at less than 50 nmol/L. The results of the meta-analysis showed a striking difference between the groups (OR: 1.47; 95%CI: 1.19–1.82; I 2 = 67%, P<0.001). As I 2 was greater than 50% ( I 2 = 67%), the random effects model was chosen to reduce heterogeneity. Three articles were included when 75 nmol/L was taken as the threshold. The results of the meta-analysis showed that there was still a striking difference in the incidence of delirium when the concentration of serum vitamin D was less than 75 nmol/L (OR: 1.54; 95%CI: 1.21–1.96; I 2 = 0%, P<0.001). A: 25nmol/L. B: 50nmol/L. C: 75nmol/L. When we took the 50 nmol / L as the cutoff value for statistical analysis, a large statistical heterogeneity was found ( I 2 = 67%). This may be due to the use of different assessment tools for delirium in the included studies, type of inpatient population, clinical setting, etc. Through subgroup analysis, different types of hospitalized patients, types of included studies, and the age difference between groups all revealed that low level of vitamin D was a risk factor for the occurrence of delirium. ( Table 3 ). Subgroup analyses for other relevant study characteristics are detailed in Table 3 . Analysis of risk factors for delirium showed that neither antipsychotic use, female gender, smoker, serum albumin content, nor BMI were risk factors for delirium in the inpatient setting ( Table 4 ). However, a meta-analysis of age showed that patients with delirium were older (WMD: 4.14; 95% CI: 0.34–7.93; I 2 = 89.7%, P = 0.03) Length of hospital stay was reported in 2 trials. Meta analysis showed that the length of hospital stay(WMD: 1.53; 95% CI: 0.55–2.51; I 2 = 77.7%, P = 0.002) was significantly different between the two groups ( Table 4 ).

In conclusion, our meta-analysis of observational studies suggests that vitamin D severe deficiency and vitamin D deficiency can increase the incidence of delirium in hospitalized patients. In addition, the study also found that 75 nmol/L may not be the threshold for the need for vitamin D supplementation preparations. The studies we included were observational, and the number of pieces of literature was small. More high-quality randomized trial studies are therefore needed to confirm this.

The current meta-analysis aims to investigate the relationship between inpatients’ delirium and low levels of vitamin D. Our systematic review was registered before the start of the literature search under the PROSPERO registration number CRD42021271347. This meta-analysis was conducted following the requirements of the MOOSE statement [ 23 ]. Inclusive criteria: 1. population: patients with serum 25(OH)D levels detected; 2. intervention vs comparator: low levels of vitamin D vs high levels of vitamin D; 3. the outcomes were the risk of delirium in patients with low levels of vitamin D and patients with higher levels of vitamin D. Represents in the form of OR, HR, etc., or enough data to calculate OR; 4. study design: observational study. Exclusion criteria: 1. those studies that could not provide enough data to calculate the incidence of delirium in vitamin D deficient patients. 2. animal studies, reviews, case reports, meeting abstracts and meta-analyses, etc. Without language restrictions, we will search the following literature databases: PubMed, EMBASE and the Cochrane Library. All the literature until September 2021 will be searched. References in the included studies were also evaluated. The keywords vitamin D, 25(OH)D, VitD, 25-hydroxyvitamin D, 25OHD, vitaminD2, vitamin D3, Hydroxycholecalciferols, hypovitaminosis D, and delirium were retrieved using medical subject headings (mesh) and text words combined. S1 Table details the complete search strategy for PubMed. Titles and abstracts of articles were independently screened by MM and NL. Full text searches were performed for eligible articles. Any discordant opinions were discussed by all authors until consensus was reached. Two authors extracted relevant information after screening the literature according to the inclusion and exclusion criteria. Author, year of publication, study type, the population included in the study, OR the value of delirium in different levels of vitamin D concentration, etc. The above data was extracted into an Excel sheet. If studies provided vitamin D concentrations of ng/ml, we performed data transformation by the formula: 1 ng/ml = 2.5 nmol/L VitD. Each enrolled article was assessed by two independent personnel using the Newcastle Ottawa scale (NOS) [ 12 ]. This scale contains study population selection, study method comparability comparison, exposure, or outcome assessment with 3-part 8 entries for a total of 9 points. The total score was 0–3 points for low-quality literature, 3–6 points for moderate-quality literature, and 6–9 points for high-quality literature. If the two investigators disagreed in the assessment of literature quality, an agreement was reached after seeking a third individual for discussion. Statistical analysis was performed using Stata 14.0 software. To evaluate the relationship between the degree and concentration of vitamin D deficiency in patients with delirium and healthy controls, raw data from each included article was extracted S2 Table . Effect sizes for dichotomous outcomes were assessed with odds ratios (OR) and 95% confidence intervals (CI). Continuous variable outcomes were presented with the weighted mean difference (WMD) and associated 95% CI. I 2 value to assess the heterogeneity of the included literature, when I 2 was less than 50%, the possibility of heterogeneity among studies was low, using the fixed effect model; Otherwise, a random effects model was used. Subgroup analysis based on possible heterogeneity factors, the robustness of results, detectable with sensitive assays.

(DOCX) Click here for additional data file. (DOCX) Click here for additional data file.