Risk Factors for Delirium in Patients with Acute Heart Failure: A Systematic Review and Meta-Analysis

Abstract

Delirium is an acute, potentially life-threatening condition characterized by altered attention, disorganized thinking, and changes in consciousness. It frequently occurs in hospitalized patients with acute heart failure (AHF). In this meta-analysis, we aimed to identify risk factors for delirium in patients with AHF (AHF-D). We evaluated all original studies on delirium occurrence in patients hospitalized for AHF. On March 11, 2024, we searched PubMed, Scopus, Ichushi, and the Cochrane Library. Data extracted included: first author’s name, publication year, inclusion/exclusion criteria, study design, delirium assessment methods, odds ratios with 95% confidence intervals, standardized mean differences, and other relevant findings. Of 2436 screened studies, 6 met eligibility criteria (3867 patients with AHF; 796 with delirium [20.6%] and 3071 without). Risk factors for AHF-D included older age; low body mass index; the use of mechanical ventilation/noninvasive positive pressure ventilation; comorbidities (previous stroke, dementia, and depression); use of antipsychotics and benzodiazepines; and laboratory findings on admission (elevated heart rate, B-type natriuretic peptide, blood urea nitrogen, serum creatinine, and low serum albumin and sodium levels). We identified 14 risk factors for AHF-D. These findings may help clinicians identify patients at high risk of developing AHF prior to delirium onset.

INTRODUCTION

Delirium is prevalent among older hospitalized patients and is an acute, potentially life-threatening condition characterized by decreased attention, disorganized thinking, and altered consciousness.^1–3) Its incidence is approximately 11–51% in the postoperative period and reaches 82% in the intensive care unit (ICU).^4,5) Delirium is associated with adverse outcomes, including prolonged hospitalization, increased mortality, and higher healthcare costs, making it a crucial factor in healthcare.^1,4,5)

Delirium is also globally prevalent among patients with acute heart failure (AHF), affecting approximately 35% of patients hospitalized with AHF.^6,7) These patients are at high risk of delirium due to advanced age, frequent emergency admissions, severe exacerbations, and invasive treatments (e.g., artificial ventilation).^6–10)

The risk factors for AHF-associated delirium (AHF-D) include male sex, older age, New York Heart Association (NYHA) class III/IV classification, a history of dementia and other conditions, the use of anxiolytic benzodiazepines (BZs), and elevated serum B-type natriuretic peptide (BNP) levels.^7,11–17)

However, these risk factors lack consistency, and it is unclear which factors are strongly associated with the development of AHF-D. Therefore, identifying the risk factors for AHF-D is a critical clinical issue; however, to date, no meta-analysis has addressed this gap.

Thus, in this study, we aimed to conduct a systematic review and meta-analysis of AHF-D to identify its risk factors and provide evidence for improving the clinical understanding of delirium in patients with AHF.

MATERIALS AND METHODS

Protocol, Search Strategy, Eligibility Criteria, and Registration

The systematic review and meta-analysis were conducted following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) protocol.¹⁸⁾ Relevant studies published up to March 11, 2024, in the PubMed, Scopus, Ichushi, and Cochrane Library databases were selected. The following text combinations were searched: (“delirium” [MeSH Terms] OR “delirium” [All Fields]) AND (“heart failure” [MeSH Terms] OR “heart failure” [All Fields]) (search details; Supplementary Material 1).

Two researchers independently screened the titles and abstracts and then reviewed the full text of the articles to determine eligibility for the meta-analysis.

The inclusion criteria were as follows: (i) patients aged ≥18 years with AHF or acute decompensated HF (ADHF); (ii) patients with delirium risk factors reported as odds ratios (ORs), continuous variables [median with interquartile range (IQR) or mean with standard deviation (S.D.)] (iii) a confirmed diagnosis of delirium in the article or use of a delirium assessment tool [Confusion Assessment Method (CAM), CAM for the ICU (CAM-ICU), Intensive Care Delirium Screening Checklist (ICDSC), or Diagnostic and Statistical Manual of Mental Disorders, 5th Edition (DSM-5); (iv) prospective or retrospective cohort studies; (v) published in peer-reviewed journals; (vi) published in English or Japanese; and (vii) studies that did not exclude patients with comorbid dementia.

The protocol was registered on PROSPERO (International Prospective Register of Systematic Reviews; CRD420251029132).

Data Extraction

An initial screening of the identified abstracts or titles was conducted to identify articles for further review. The second screening was based on a full-text review to identify studies that met the inclusion criteria. The literature screening, data extraction, and cross-checking were performed independently by 2 researchers using the literature review tool “Rayyan.”¹⁹⁾ Disagreements were resolved through consultation and discussion with a third researcher.

The following variables were collected from the articles: author’s family name, year of publication, study design, country of origin, sample size, assessment tool, and number of patients with delirium. The measurement of each variable was confirmed according to the corresponding criteria of the included studies, and delirium was diagnosed based on the criteria of the included studies.

Means and S.Ds. of continuous variables were extracted. However, some studies only reported the median and IQR of the variables; therefore, these data were converted to means and S.Ds. using a conversion formula, following previous studies,⁸⁾ and used in the analysis.²⁰⁾

Ethical Considerations and Quality of Included Studies

As a systematic review that did not involve human participants or personal data, this study did not require ethics committee approval. Two researchers independently assessed the quality of the included studies using the Newcastle–Ottawa Scale.²¹⁾ Disagreements were resolved through discussions with a third researcher. The NOS is an effective method for assessing the quality of systematic reviews of observational studies. The assessment includes three areas—selection of study participants (4 points), control of confounding factors in the study cohort (2 points), and judgment of outcome events (3 points). There are 8 items on the scale, with a total score of 9.0. A score between 7.0 and 9.0 indicates high quality, whereas that between 4.0 and 6.0 indicates medium quality.

Statistical Analyses

Binary variables using pooled ORs with 95% CIs were analyzed and calculated via the Mantel–Haenszel method. For continuous variables, we determined standardized mean differences (SMDs) with 95% CIs using inverse variance weighting. Statistical significance was set at p < 0.05 for both ORs and SMDs. For continuous variables, we computed weighted average thresholds (WATs) by multiplying each study’s mean value (as shown in the forest plot) by its proportional weight and summing these values across studies.

Heterogeneity was assessed using the Cochran chi-squared test (Q-test) and the I² test. The p-value for heterogeneity in the Q-test was denoted as “p-het.” A p-het < 0.05 suggested statistically significant heterogeneity. The I² values of 25, 50, and 75% represented low, moderate, and high heterogeneity, respectively.²²⁾ Based on previous research,⁸⁾ a random effects model was used when p-het <0.1 and I² ≥75% indicated significant heterogeneity between studies; otherwise, a fixed effects model was used. If heterogeneity was greater than 75% and there were more than 4 studies, a sensitivity analysis was performed. Publication bias was assessed using funnel plots (Supplementary Material 2).

All statistical analyses were performed using EZR ver1.68,²³⁾ a more precise, modified version of the R (R Foundation for Statistical Computing, Vienna, Austria) command designed to add the statistical functions frequently used in biostatistics.

RESULTS

General Characteristics of the Included Studies

Figure 1 shows a flow diagram of the study screening process. Notably, 2163 abstracts of the 2436 identified articles were screened after removing duplicates (n = 273). Furthermore, 2143 articles were excluded after screening titles and abstracts. A full-text review of the remaining 20 studies resulted in the exclusion of 14 articles and the inclusion of 6 studies in the meta-analysis. There were 3 retrospective cohort studies^12,13,16) and 3 prospective cohort studies^14,15,17); 4 were conducted in Japan,^12–15) 1 in the U.S.A.,¹⁶⁾ and 1 in Spain.¹⁷⁾ Table 1 summarizes the characteristics of these studies. The quality of the 6 included papers^12–15) was assessed using the NOS, and all studies had an NOS score of ≥8, indicating a low risk of bias (Table 2).

Fig. 1. PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analysis) Flowchart

Table 1. General Characteristics of the Included Studies

Study	Country	Study design	HF type	Sample size	Number of patients		Age (year)		Male (%)	Diagnostic criteria	Follow-up (days)	Cognitive function test	NOS
					Delirium	Non-delirium	Delirium	Non-delirium
Kawada et al. (2021)12)	Japan	Retrospective cohort	ADHF	650	59	591	81.7 ± 9.1*	74.5 ± 12.8*	56.3	CAM	365	MMSE	9
Aikawa et al. (2024)13)	Japan	Retrospective cohort	AHF	1555	406	1149	84 (78–89)**	78 (69–84)**	56.8	ICDSC CAM-ICU	365	Not stated	9
Iwata et al. (2020)14)	Japan	Prospective cohort	AHF	408	109	299	85 (79–90)**	79 (70–85)**	53.1	CAM-ICU	365	MMSE	9
Pak et al. (2020)15)	Japan	Prospective cohort	ADHF	132	36	96	82 (75–86)**	85 (83–91)**	51.5	DSM-5	90	Not stated	8
Uthamalingam et al. (2011)16)	U.S.A.	Retrospective cohort	ADHF	883	151	732	83 ± 8*	79 ± 8*	48	CAM	90	AMT	8
Rizzi et al. (2015)17)	Spain	Prospective cohort	ADHF	239	35	204	86.8 ± 6.6*	80.8 ± 9.6*	38.9	CAM	365	MMSE	8

^*Mean ± S.D. ^**Median (IQR). ADHF: acute decompensated heart failure; AHF: acute heart failure; AMT: Abbreviated Mental Test; CAM: Confusion Assessment Method; CAM-ICU: CAM for the intensive care unit; DSM - 5: Diagnostic and Statistical Manual of Mental Disorders, 5th Edition; HF: heart failure; ICDSC: Intensive Care Delirium Screening Checklist; IQR: interquartile range; MMSE: Mini-Mental State Examination; NOS: The Newcastle–Ottawa scale; S.D.: standard deviation (details are in Table 2).

Table 2. The Newcastle–Ottawa Scale

Study	Cohort selection (A–D)				Comparability	Outcome (F–H)			Total score
	A	B	C	D	E	F	G	H
Kawada et al. (2021)12)	1	1	1	1	2	1	1	1	9
Aikawa et al. (2024)13)	1	1	1	1	2	1	1	1	9
Iwata et al. (2020)14)	1	1	1	1	2	1	1	1	9
Pak et al. (2020)15)	1	1	1	1	2	1	0	1	8
Uthamalingam et al. (2011)16)	1	1	1	1	2	1	0	1	8
Rizzi et al. (2015)17)	1	1	1	1	2	1	0	1	8

(A) Representativeness of the exposed cohort, (B) representativeness of the nonexposed cohort, (C) ascertainment of exposure, (D) outcome of interest was not present at the start of the study, (E) comparability of the exposed and nonexposed cohorts, (F) quality of outcome assessment, (G) follow-up long enough for outcomes to occur, and (H) adequacy of follow-up.

Risk Factors for AHF-D

A total of 16 factors were analyzed, and 14 were statistically significant and could induce delirium (Table 3).

Table 3. Results of the Meta-Analysis

Variables	Number of studies	Total sample			WAT		Heterogeneity* and statistical method
			OR/SMD (95% CI), p-value		Experimental	Control	I2 (%)	Q-test (p-het)	Estimation method
Patient background
Age (years)	12, 13, 14, 15, 16, 17	3867	SMD	0.61 (0.53, 0.69), <0.01	83.9	77.8	0.0	0.84	IV, fixed
Male (vs. female)	12, 13, 14, 15, 16, 17	3867	OR	0.87 (0.74, 1.02), 0.08			13	0.33	M–H, fixed
BMI (kg/m2)	12, 13, 14, 15	2745	SMD	−0.41 (−0.50, −0.32), <0.01	20.2	21.7	50	0.11	IV, fixed
Pathology and clinical intervention
NYHA classification III/IV	12, 13, 14, 16	3496	OR	1.64 (0.87, 3.08), 0.13			78	<0.01	M–H, random
Sensitivity analysis; NYHA classification III/IV	12, 16	1533	OR	2.90 (1.92, 4.40), <0.01			61	0.11	M–H, fixed
MV/NPPV	12, 13, 15	2337	OR	3.08 (1.32, 7.20), <0.01			87	<0.01	M–H, random
Comorbidities
Previous stroke	14, 15, 16, 17	1662	OR	1.96 (1.38, 2.77), <0.01			37	0.19	M–H, fixed
Dementia	12, 13, 14, 15, 16, 17	3867	OR	5.64 (3.04, 10.4), <0.01			88	<0.01	M–H, random
Sensitivity analysis; dementia	12, 13, 14, 15, 17	2984	OR	9.00 (7.25, 11.18), <0.01			0.0	0.48	M–H, fixed
Depression	12, 16, 17	1772	OR	1.72 (1.23, 2.41), <0.01			61	0.08	M–H, fixed
Concomitant medications
Antipsychotics	13, 14, 15	2095	OR	11.9 (4.27, 33.0), < 0.01			77	0.01	M–H, random
Benzodiazepines	12, 13, 14, 15	2745	OR	1.26 (1.02, 1.56), 0.03			66	0.03	M–H, fixed
Laboratory findings on admission
Heart rate (bpm)	13, 14, 15, 16	2978	SMD	0.13 (0.05, 0.22), <0.01	91.8	88.6	54	0.09	IV, fixed
BNP (pg/mL)	12, 13, 14, 15, 16	3628	SMD	0.40 (0.32, 0.49), <0.01	1097	772	74	<0.01	IV, fixed
Albumin (g/dL)	13, 14, 16	2846	SMD	−0.49 (−0.57, −0.40), <0.01	3.30	3.52	29	0.24	IV, fixed
BUN (mg/dL)	13, 14, 16, 17	3085	SMD	0.20 (0.11, 0.28), <0.01	29.7	27.2	27	0.25	IV, fixed
Creatinine (mg/dL)	13, 14, 16, 17	3085	SMD	0.21 (0.13, 0.30), <0.01	1.39	1.26	15	0.31	IV, fixed
Sodium (mEq/mL)	13, 14, 16, 17	3085	SMD	−0.24 (−0.32, −0.15), <0.01	138.6	139.5	53	0.09	IV, fixed

The standardized mean difference (SMD) of the continuous variables and the corresponding 95% CI were estimated using the IV method. The pooled OR and corresponding 95% CIs for binary variables were estimated using M–H method. The random effects model was used when I² was ≥75% and p-het was <0.1. Otherwise, a fixed effects model was used. *Heterogeneity is tested using both the Cochran chi-square test (Q-test) and I² test. BMI: body mass index; BNP: b-type natriuretic peptide; BUN: blood urea nitrogen; CI: confidence interval; IV: inverse variance; M–H: Mantel–Haenszel; MV: mechanical ventilation; NPPV: non-invasive positive pressure ventilation; NYHA: New York Heart Association; OR: odds ratio; p-het: p-value for heterogeneity in Q-test; SMD: standardized mean difference; WAT: weighted average threshold.

Patient Background

Six studies^12–17) reported patient ages. The estimated SMD for age was 0.61 (95% CI: 0.53, 0.69, p < 0.01), with low heterogeneity (I² = 0.0%, p-het = 0.84), indicating older age as a significant risk factor. The WAT for age was 83.9 years in the delirium group and 77.8 years in the control group. In addition, all 6 studies^12–17) reported the sex of patients. The pooled OR for male vs. females was 0.87 (95% CI: 0.74, 1.02, p = 0.08), with low heterogeneity (I² = 13%, p-het = 0.33), indicating no significant association between male sex and delirium risk. Four studies^12–15) included body mass index (BMI) data. The analysis revealed an SMD of −0.41 (95% CI = −0.50, −0.32, p < 0.01) with low heterogeneity (I² = 50%, p-het = 0.11), establishing low BMI as a significant risk factor. The WAT values were 20.2 kg/m² for the delirium group and 21.7 kg/m² for controls (Table 3 and Figs. 2A–2C).

Fig. 2. Forest Plot of Individual Factors Associated with AHF-D

Patient background: (A) age (years), (B) male (vs. female), (C) BMI; pathology and clinical intervention: (D) NYHA classification III/IV, (E) MV/NPPV; comorbidities: (F) previous stroke, (G) dementia, (H) depression; concomitant medications: (I) antipsychotics, (J) benzodiazepines; laboratory findings at admission: (K) HR, (L) BNP, (M) albumin, (N) BUN, (O) creatinine, (P) sodium. Results are presented as ORs or SMDs, with corresponding 95% CIs. Heterogeneity across studies was assessed using the I² statistic and Cochran’s Q-test (p-heterogeneity). AHF-D: acute heart failure-related delirium; BMI: body mass index; BNP: B-type natriuretic peptide; BUN: blood urea nitrogen; CI: confidence interval; HR: heart rate; MV: mechanical ventilation; NPPV: noninvasive positive pressure ventilation; NYHA: New York Heart Association; OR: odds ratio; S.D: standard deviation; SMD: standardized mean difference; WAT: weighted average threshold.

Fig. 2. Forest Plot of Individual Factors Associated with AHF-D

Patient background: (A) age (years), (B) male (vs. female), (C) BMI; pathology and clinical intervention: (D) NYHA classification III/IV, (E) MV/NPPV; comorbidities: (F) previous stroke, (G) dementia, (H) depression; concomitant medications: (I) antipsychotics, (J) benzodiazepines; laboratory findings at admission: (K) HR, (L) BNP, (M) albumin, (N) BUN, (O) creatinine, (P) sodium. Results are presented as ORs or SMDs, with corresponding 95% CIs. Heterogeneity across studies was assessed using the I² statistic and Cochran’s Q-test (p-heterogeneity). AHF-D: acute heart failure-related delirium; BMI: body mass index; BNP: B-type natriuretic peptide; BUN: blood urea nitrogen; CI: confidence interval; HR: heart rate; MV: mechanical ventilation; NPPV: noninvasive positive pressure ventilation; NYHA: New York Heart Association; OR: odds ratio; S.D: standard deviation; SMD: standardized mean difference; WAT: weighted average threshold.

Fig. 2. Forest Plot of Individual Factors Associated with AHF-D

Patient background: (A) age (years), (B) male (vs. female), (C) BMI; pathology and clinical intervention: (D) NYHA classification III/IV, (E) MV/NPPV; comorbidities: (F) previous stroke, (G) dementia, (H) depression; concomitant medications: (I) antipsychotics, (J) benzodiazepines; laboratory findings at admission: (K) HR, (L) BNP, (M) albumin, (N) BUN, (O) creatinine, (P) sodium. Results are presented as ORs or SMDs, with corresponding 95% CIs. Heterogeneity across studies was assessed using the I² statistic and Cochran’s Q-test (p-heterogeneity). AHF-D: acute heart failure-related delirium; BMI: body mass index; BNP: B-type natriuretic peptide; BUN: blood urea nitrogen; CI: confidence interval; HR: heart rate; MV: mechanical ventilation; NPPV: noninvasive positive pressure ventilation; NYHA: New York Heart Association; OR: odds ratio; S.D: standard deviation; SMD: standardized mean difference; WAT: weighted average threshold.

Pathology and Clinical Intervention

Four studies^12–14,16) examined NYHA classification III/IV as a potential risk factor. The pooled OR was 1.64 (95% CI: 0.87, 3.08, p = 0.13), with high heterogeneity (I² = 78%, p-het < 0.01), indicating NYHA classification III/IV was not a significant risk factor. Three studies^12,13,15) evaluated mechanical ventilation (MV) and noninvasive positive pressure ventilation (NPPV). The pooled OR was 3.08 (95% CI: 1.32, 7.20, p < 0.01), with high heterogeneity (I² = 87%, p-het < 0.01) (Table 3 and Figs. 2D, 2E).

Comorbidities

Four studies^14–17) reported on the incidence of previous stroke in patients. The pooled OR was 1.96 (95% CI: 1.38, 2.77, p < 0.01), with moderate heterogeneity (I² = 37%, p-het = 0.19). Furthermore, 6 studies^12–17) reported the prevalence of dementia in patients. The pooled OR was 5.64 (95% CI: 3.04, 10.4, p < 0.01), with high heterogeneity (I² = 88%, p-het < 0.01). Three studies^12,16,17) reported the prevalence of depression. The pooled OR was 1.72 (95% CI: 1.23, 2.41, p < 0.01), with high heterogeneity (I² = 61%, p-het = 0.08) (Table 3 and Figs. 2F–2H).

Concomitant Medications

Three studies^13–15) reported the use of antipsychotics in patients. The pooled OR was 11.9 (95% CI: 4.27, 33.0, p < 0.01), with high heterogeneity (I² = 77%, p-het = 0.01). Four studies^12–15) reported on the use of BZs. The pooled OR was 1.26 (95% CI: 1.02, 1.56, p = 0.03), with moderate heterogeneity (I² = 66%, p-het = 0.03) (Table 3 and Figs. 2I, 2J).

Laboratory Findings at Admission

Four studies^13–16) reported heart rates (HRs). The estimated SMD was 0.13 (95% CI: 0.05, 0.22, p < 0.01), with moderate heterogeneity (I² = 54%, p-het = 0.09), suggesting high HR as a risk factor. The WAT for HR was 91.8 bpm in the delirium group and 88.6 bpm in the control group. Furthermore, 5 studies^12–16) reported the serum levels of BNPs. The estimated SMD was 0.40 (95% CI: 0.32, 0.49, p < 0.01), with moderate heterogeneity (I² = 74%, p-het< 0.01), indicating high serum BNP levels as a risk factor. The WAT for BNP was 1097 pg/mL in the delirium group and 772 pg/mL in the control group. Three studies^13,14,16) reported serum albumin levels. The estimated SMD was −0.49 (95% CI: −0.57, −0.40, p < 0.01), with low heterogeneity (I² = 29%, p-het = 0.24), suggesting low serum albumin as a risk factor. The WAT for albumin was 3.30 g/dL in the delirium group and 3.52 g/dL in the control group. Four studies^13,14,16,17) reported blood urea nitrogen (BUN), serum creatinine, and sodium concentration levels. The estimated SMD for high BUN was 0.20 (95% CI: 0.11, 0.28, p < 0.01), with low heterogeneity (I² = 27%, p-het = 0.25). The WAT for BUN was 29.7 mg/dL in the delirium group and 27.2 mg/dL in the control group. The estimated SMD for high serum creatinine was 0.21 (95% CI: 0.13, 0.30, p < 0.01), with low heterogeneity (I² = 15%, p-het = 0.31). The WAT for creatinine was 1.39 mg/dL in the delirium group and 1.26 mg/dL in the control group. The estimated SMD for low sodium was −0.24 (95% CI: −0.32, −0.15, p < 0.01), with moderate heterogeneity (I² = 53%, p-het = 0.09). The WAT for sodium was 138.6 mEq/mL in the delirium group and 139.5 mEq/mL in the control group (Table 3 and Figs. 2K–2P).

Sensitivity Analysis

We performed sensitivity analyses for outcomes with substantial heterogeneity (I² > 75%) from 4 or more studies. For NYHA classification, 2 ICU-exclusive studies^13,14) likely contributed to high heterogeneity due to the exclusive inclusion of more severely ill populations. Therefore, we conducted stratified meta-analyses separating ICU-exclusive studies from others. The ICU-focused studies (n = 2) suggested that NYHA class III/IV may be a significant risk factor for delirium, with a pooled OR of 2.90 (95% CI: 1.92, 4.40, p < 0.01) and moderate heterogeneity (I² = 61%, p-het = 0.11). However, since only 2 studies were available for this subgroup, we suggest that drawing a definitive quantitative conclusion is premature, and the findings should be interpreted with caution. For dementia analysis, we identified that Uthamalingam et al.¹⁶⁾ exclusively used the Abbreviated Mental Test (AMT) rather than the Mini-Mental State Examination (MMSE), potentially contributing to heterogeneity. After excluding this study,¹⁶⁾ the reanalysis showed reduced heterogeneity and a significant association between dementia and delirium (pooled OR = 9.00; 95% CI: 7.25, 11.2, p < 0.01; I² = 0%, p-het = 0.48) (Table 3 and Figs. 3A–3C).

Fig. 3. Forest Plot of Individual Factors Associated with AHF-D

Sensitivity analysis of highly heterogeneous factors: (A) NYHA classification III/IV (excluding study by Iwata et al.¹⁴⁾ and Aikawa et al.¹³⁾), (B) NYHA classification III/IV (excluding study by Uthamalingam et al.¹⁶⁾ and Kawada et al.¹²⁾), and (C) dementia (excluding study by Uthamalingam et al.¹⁶⁾). Results were summarized with OR and corresponding 95% CI. I² statistics and Q-test (p-heterogeneity) were used to assess heterogeneity among studies. AHF-D: acute heart failure-related delirium; CI: confidence interval; NYHA: New York Heart Association; OR: odds ratio.

DISCUSSION

To our knowledge, this study is the first systematic review and meta-analysis to provide evidence on risk factors for AHF-D, which is prevalent among older patients. Our results showed that the risk factors for AHF-D included older age, low BMI, the use of MV/NPPV, comorbidities (previous stroke, dementia, and depression), the use of various medications (antipsychotics and BZs), and laboratory findings on admission (high HR, high BNP levels, low serum albumin, high BUN, high serum creatinine, and low sodium).

Delirium is an acute neuropsychiatric disorder that is characterized by transient cognitive impairment. Its pathophysiology is multifaceted and involves multiple interrelated mechanisms, including cerebral hypoperfusion, neuroinflammation, neurotransmitter dysregulation, and sleep–wake rhythm disorders.^24–26)

In AHF, reduced cerebral blood flow due to low cardiac output, a systemic inflammatory response, and poor oxygenation are thought to affect brain function, contributing to delirium pathogenesis.^27–29) Several analyses of delirium risk factors have been conducted in patients with AHF. In this study, we conducted a systematic review and meta-analysis of these reports.

Consistent with previous reports, older age emerged as a risk factor. Patients with HF are typically older and often frail, malnourished, or overweight, with low albumin levels or BMI, respectively, showing a clear correlation with delirium development.¹⁵⁾ The results of the WAT suggest heightened delirium risk for patients aged >83.9 years or with a BMI <20.2 kg/m². Delirium is reported to occur more frequently in males; however, our meta-analysis revealed no significant sex-based differences among patients with AHF. The underlying reasons for this discrepancy remain unclear and warrant further investigation. It is possible that, in the AHF population, the occurrence of delirium does not substantially differ between sexes.

Both previous studies and our meta-analysis identified MV/NPPV as significant risk factors for delirium.^12,13,15) While these interventions are life-saving, they induce substantial physiological stress and trigger systemic inflammatory responses that affect the central nervous system, resulting in the activation of immune cells such as microglia in the brain, the release of various inflammatory cytokines such as tumor necrosis factor-α (TNF-α) and interleukin-1β (IL-1β), and neurological dysfunction and cerebral blood flow interruption, which may lead to delirium.^{8–10,30–33)} Two factors are critical when elucidating the pathophysiology of delirium: direct brain damage and an abnormal stress response.^24,34,35) Therefore, appropriate monitoring is considered necessary to manage this highly invasive condition that is prone to abnormal stress responses.

Uthamalingam et al. identified NYHA classification III/IV as a risk factor for delirium in patients with ADHF.¹⁶⁾ This association is biologically plausible given the generally unstable hemodynamic and respiratory status. In our meta-analysis, 2 studies that included both general ward and ICU patients^13,14) yielded a pooled OR of 2.90 (95% CI: 1.92–4.40, p < 0.01), suggesting that NYHA class III/IV may be a significant risk factor for delirium. In contrast, 2 studies that included only ICU patients did not identify NYHA class III/IV as a significant risk factor. This discrepancy may be explained by the fact that most ICU patients are already classified as having severe heart failure (NYHA class III/IV), making this classification less discriminative for predicting the onset of delirium in this population. However, as each of these subgroup analyses was based on only 2 studies, the findings should be interpreted with caution.

Our meta-analysis estimated that patients with a history of stroke, dementia, or depression were more likely to develop delirium. The analysis of dementia demonstrated particularly high heterogeneity. Uthamalingam et al.’s study utilized the AMT, while other studies employed the MMSE. Studies have reported that the AMT has lower sensitivity and specificity in detecting cognitive impairment compared with the MMSE.³⁶⁾ This may have caused differences in the severity of dementia. To address this heterogeneity, we performed a sensitivity analysis excluding the Uthamalingam et al.¹⁶⁾ study, resulting in 0% heterogeneity. While delirium and dementia are pathologies that should be clearly distinguished from one another, neuronal damage and cell death are likely to have occurred in the brains of patients with incidents of previous stroke or dementia, and this type of damage is consistent with the mechanism of delirium onset.²⁴⁾

Delirium in patients with depression is thought to result from a multifactorial interplay involving neurotransmitter imbalances, neuroinflammation, and blood–brain barrier dysfunction. The critical pathway involves inflammatory cytokines, including IL-6, TNF-α, and IFN-γ, which become elevated due to systemic stressors such as infection, surgery, or chronic disease. These cytokines activate indoleamine 2,3-dioxygenase, shifting tryptophan metabolism away from serotonin synthesis toward the kynurenine pathway. This metabolic shift reduces serotonin availability while increasing the production of neurotoxic metabolites like quinolinic acid, ultimately impairing cognition and contributing to delirium onset. In depression, a chronic low-grade inflammatory state further predisposes individuals to this vulnerability. Understanding these shared mechanisms between depression and delirium may inform future prevention and treatment strategies for vulnerable psychiatric populations.^37–40)

Our meta-analysis identified antipsychotic medications as a risk factor for AHF-D. Delirium is caused by an excess of dopaminergic activity in the brain.⁴¹⁾ Antipsychotics, commonly used to treat schizophrenia, have a complex relationship with delirium; however, they are effective in managing the agitation and psychiatric symptoms associated with delirium by blocking dopamine receptors. However, they may also increase the risk of delirium as a side effect due to their anticholinergic effects.⁴²⁾ Anticholinergic delirium arises from central cholinergic deficiency, primarily due to muscarinic M1 receptor antagonism, leading to impairments in attention, memory, and perception. Older individuals are particularly susceptible due to age-related cholinergic function decline.⁴³⁾ These findings suggest that the use of antipsychotics may be a risk factor in patients with AHF.

In this study, BZ use was estimated to be a risk factor for AHF-D. BZs act on gamma-aminobutyric acid-A (GABA-A) receptors to enhance inhibitory neurotransmission. However, excessive GABA activation can lead to imbalances in other neurotransmitters, such as dopamine, serotonin, acetylcholine, norepinephrine, and glutamate, which can result in delirium. Overstimulation of the GABAergic system is thought to cause sensory overload and confusion by reducing cortical–striatal glutamatergic tone and impairing thalamic filtering function.⁴⁴⁾ BZs are fat-soluble drugs, and their half-lives are prolonged in older adults because they accumulate in lipid tissues. Therefore, BZs can cause delirium because of their longer duration of action and the increased sensitivity to sedative hypnotics in older adults.⁴⁵⁾ The present results confirm our previous findings.¹²⁾

A high HR was shown to be a possible risk factor for developing AHF-D. Previous studies have shown that the development of delirium due to overstress is correlated with an increased HR. Ernst et al. stated that HR variability (HRV) is a method of assessing the autonomic nervous system and may reflect central brain status. They also reported that delirium in patients with a hip fracture was associated with increased HRV.⁴⁶⁾ Preoperative HRV analysis can be used as an indicator to predict perioperative delirium in patients undergoing esophageal surgery.⁴⁷⁾

High BNP levels were estimated to be a risk factor for AHF-D. The level of BNPs is an independent risk factor for delirium in patients admitted with AHF.^7,16) BNPs are cardiac neurohormones that are secreted by the ventricles, particularly in response to increased ventricular volume expansion and pressure. The BNP level is one of the strongest risk predictors of mortality and AHF hospitalization⁴⁸⁾; therefore, it is recommended as a standard biomarker for AHF diagnosis. However, in this study, it is also shown to be a predictor of AHF-D expression.

High BUN and serum creatinine levels were estimated to be risk factors. Aikawa et al. reported that, in patients with AHF admitted to the cardiac ICU, BUN was higher in the group with prolonged delirium than in the group with resolved delirium.¹³⁾ A meta-analysis of 4 studies^13,14,16,17) reporting the use of serum creatinine to assess renal function estimated that a high serum creatinine level is a risk factor for AHF-D. A low estimated glomerular filtration rate is associated with the occurrence of delirium.¹²⁾ Fluid retention and hemodynamic compromise due to poor renal function may lead to cerebral hypoperfusion, which may correlate with the development of delirium.^27,28) Renal function is also associated with AHF status and may reflect the pathophysiology of AHF.

Our meta-analysis suggests that hyponatremia is a risk factor for AHF-D. The mild symptoms of hyponatremia include nausea, vomiting, weakness, headache, and mild neurocognitive impairment, whereas the severe symptoms include delirium, confusion, disorientation, ataxia, seizures, and, rarely, brain herniation or death. Furthermore, when assessing patients, clinicians should classify them based on their fluid volume status.^48–50) Therefore, fluid management in patients with AHF is essential, and sodium concentration is a critical indicator.

Our meta-analysis has some limitations. First, the small number of included studies limited our ability to fully assess heterogeneity. In addition, it was difficult to visually assess the funnel plot asymmetry, affecting our assessment of publication bias.⁵¹⁾ Second, 4 of the 6 included studies were from Japan, which may limit the generalizability of our findings to other populations. Third, assessment methods for dementia and delirium were not standardized among the included studies, potentially leading to bias due to differences in diagnostic criteria. Fourth, regarding biochemical tests, there were no differences in the measurement methods described in the guidelines between Japan, the United States, and Spain.^52–54) However, there are multiple approved measurement methods, and the methods used by laboratories may differ. These measurement methods were not specified in the paper, making their impact unclear.

CONCLUSION

In this meta-analysis, we revealed patient background (older age and low BMI), the use of MV/NPPV, comorbidities (previous stroke, dementia, and depression), medications used (antipsychotics and BZs), and laboratory findings on admission (high HR and high BNP, low serum albumin, high BUN, high serum creatinine, and low sodium levels) as risk factors for AHF-D. Therefore, upon admission for AHF exacerbation, these factors should be promptly identified and appropriate measures taken to prevent AHF-D development. We hope that this study will help to prevent delirium during hospitalization for AHF exacerbations.

Acknowledgments

We want to thank all collaborators. No funding was received for this study.

Author Contributions

Tomoaki Ishida contributed to the conception and design of the study. Tetsushi Kawazoe, Tomoaki Ishida, and Kohei Jobu performed the literature screening, data extraction, and quality assessment of the study. Tetsushi Kawazoe wrote the first draft of the manuscript. Tomoaki Ishida, Kohei Jobu, Kei Kawada, Shumpei Morisawa, Junko Tomida, Naomi Iihara, Yoichi Kawasaki, and Yukihiro Hamada commented on the previous draft. All authors have read and approved the final manuscript.

Conflict of Interest

The authors declare no conflict of interest.

Data Availability

If necessary, all data can be obtained from the corresponding author.

Supplementary Materials

This article contains supplementary materials.

REFERENCES

• 1)  Cole MG, Primeau FJ. Prognosis of delirium in elderly hospital patients. CMAJ, 149, 41–46 (1993).
• 2)  Ely EW, Margolin R, Francis J, May L, Truman B, Dittus R, Speroff T, Gautam S, Bernard GR, Inouye SK. Evaluation of delirium in critically ill patients: validation of the Confusion Assessment Method for the intensive care unit (CAM-ICU). Crit. Care Med., 29, 1370–1379 (2001).
• 3)  Inouye SK, van Dyck CH, Alessi CA, Balkin S, Siegal AP, Horwitz RI. Clarifying confusion: the confusion assessment method: a new method for detection of delirium. Ann. Intern. Med., 113, 941–948 (1990).
• 4)  Inouye SK, Robinson T, Blaum C, et al. Postoperative delirium in older adults: best practice statement from the American geriatrics society. J. Am. Coll. Surg., 220, 136–148e1 (2015).
• 5)  Hshieh TT, Inouye SK, Oh ES. Delirium in the elderly. Psychiatr. Clin. North Am., 41, 1–17 (2018).
• 6)  Correale M, Altamura M, Carnevale R, Tricarico L, Malerba S, Gallotta AM, Altamura C, Perrella M, Di Biase M, Bellomo A, Brunetti ND. Delirium in heart failure. Heart Fail. Rev., 25, 713–723 (2020).
• 7)  Honda S, Nagai T, Sugano Y, Okada A, Asaumi Y, Aiba T, Noguchi T, Kusano K, Ogawa H, Yasuda S, Anzai T. Prevalence, determinants, and prognostic significance of delirium in patients with acute heart failure. Int. J. Cardiol., 222, 521–527 (2016).
• 8)  Wu J, Yin Y, Jin M, Li B. The risk factors for postoperative delirium in adult patients after hip fracture surgery: a systematic review and meta-analysis. Int. J. Geriatr. Psychiatry, 36, 3–14 (2021).
• 9)  Robinson TN, Raeburn CD, Tran ZV, Angles EM, Brenner LA, Moss M. Postoperative delirium in the elderly: risk factors and outcomes. Ann. Surg., 249, 173–178 (2009).
• 10)  Elie M, Rousseau F, Cole M, Primeau F, McCusker J, Bellavance F. Prevalence and detection of delirium in elderly emergency department patients. CMAJ., 163, 977–981 (2000).
• 11)  Niu Z, Zhou J, Li Y. Prognostic significance of delirium in patients with heart failure: a systematic review and meta-analysis. Front. Cardiovasc. Med., 10, 1217965 (2023).
• 12)  Kawada K, Fukuda H, Kubo T, Ohta T, Ishida T, Morisawa S, Kawazoe T, Okamoto M, Fujita H, Jobu K, Morita Y, Ueba T, Kitaoka H, Miyamura M. Added value of anxiolytic benzodiazepines in predictive models on severe delirium in patients with acute decompensated heart failure: a retrospective analysis. PLOS ONE, 16, e0250372 (2021).
• 13)  Aikawa Y, Ogata S, Honda S, Nagai T, Murata S, Morii I, Anzai T, Nishimura K, Noguchi T. Prolonged delirium during hospitalization is associated with worse long-term and short-term outcomes in patients with acute heart failure. Int. J. Cardiol., 399, 131776 (2024).
• 14)  Iwata E, Kondo T, Kato T, Okumura T, Nishiyama I, Kazama S, Ishihara T, Kondo S, Hiraiwa H, Tsuda T, Ito M, Aoyama M, Tanimura D, Awaji Y, Unno K, Murohara T. Prognostic value of delirium in patients with acute heart failure in the intensive care unit. Can. J. Cardiol., 36, 1649–1657 (2020).
• 15)  Pak M, Hara M, Miura S, Furuya M, Tamaki M, Okada T, Watanabe N, Endo A, Tanabe K. Delirium is associated with high mortality in older adult patients with acute decompensated heart failure. BMC Geriatr., 20, 524 (2020).
• 16)  Uthamalingam S, Gurm GS, Daley M, Flynn J, Capodilupo R. Usefulness of acute delirium as a predictor of adverse outcomes in patients >65 years of age with acute decompensated heart failure. Am. J. Cardiol., 108, 402–408 (2011).
• 17)  Rizzi MA, Torres Bonafonte OH, Alquezar A, Herrera Mateo S, Piñera P, Puig M, Benito S, Ruiz D. Prognostic value and risk factors of delirium in emergency patients with decompensated heart failure. J. Am. Med. Dir. Assoc., 16, 799.e1–799.e6 (2015).
• 18)  Page MJ, McKenzie JE, Bossuyt PM, et al. The PRISMA, 2020 statement: an updated guideline for reporting systematic reviews. J. Clin. Epidemiol., 134, 178–189 (2021).
• 19)  Ouzzani M, Hammady H, Fedorowicz Z, Elmagarmid A. Rayyan—a web and mobile app for systematic reviews. Syst. Rev., 5, 210 (2016).
• 20)  Wan X, Wang W, Liu J, Tong T. Estimating the sample mean and standard deviation from the sample size, median, range and/or interquartile range. BMC Med. Res. Methodol., 14, 135 (2014).
• 21)  Stang A. Critical evaluation of the Newcastle-Ottawa scale for the assessment of the quality of nonrandomized studies in meta-analyses. Eur. J. Epidemiol., 25, 603–605 (2010).
• 22)  Higgins JPT, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta-analyses. BMJ, 327, 557–560 (2003).
• 23)  Kanda Y. Investigation of the freely available easy-to-use software ‘EZR’ for medical statistics. Bone Marrow Transplant., 48, 452–458 (2013).
• 24)  Maldonado JR. Neuropathogenesis of delirium: review of current etiologic theories and common pathways. Am. J. Geriatr. Psychiatry., 21, 1190–1222 (2013).
• 25)  Subramaniyan S, Terrando N. Neuroinflammation and perioperative neurocognitive disorders. Anesth. Analg., 128, 781–788 (2019).
• 26)  Terzaghi M, Sartori I, Rustioni V, Manni R. Sleep disorders and acute nocturnal delirium in the elderly: a comorbidity not to be overlooked. Eur. J. Intern. Med., 25, 350–355 (2014).
• 27)  Fong TG, Bogardus ST Jr, Daftary A, Auerbach E, Blumenfeld H, Modur S, Leo-Summers L, Seibyl J, Inouye SK. Cerebral perfusion changes in older delirious patients using 99mTc HMPAO SPECT. J. Gerontol. A Biol. Sci. Med. Sci., 61, 1294–1299 (2006).
• 28)  Yokota H, Ogawa S, Kurokawa A, Yamamoto Y. Regional cerebral blood flow in delirium patients. Psychiatry Clin. Neurosci., 57, 337–339 (2003).
• 29)  Boulet J, Sridhar VS, Bouabdallaoui N, Tardif JC, White M. Inflammation in heart failure: pathophysiology and therapeutic strategies. Inflamm. Res., 73, 709–723 (2024).
• 30)  Zhu C, Wang B, Yin J, Xue Q, Gao S, Xing L, Wang H, Liu W, Liu X. Risk factors for postoperative delirium after spinal surgery: a systematic review and meta-analysis. Aging Clin. Exp. Res., 32, 1417–1434 (2020).
• 31)  Gosselt AN, Slooter AJ, Boere PR, Zaal IJ. Risk factors for delirium after on-pump cardiac surgery: a systematic review. Crit. Care, 19, 346 (2015).
• 32)  Lee SH, Lim SW. Risk factors for postoperative delirium after colorectal surgery: a systematic review and meta-analysis. Int. J. Colorectal Dis., 35, 433–444 (2020).
• 33)  Shi C, Yang C, Gao R, Yuan W. Risk factors for delirium after spinal surgery: a meta-analysis. World Neurosurg., 84, 1466–1472 (2015).
• 34)  MacLullich AMJ, Ferguson KJ, Miller T, de Rooij SEJA, Cunningham C. Unravelling the pathophysiology of delirium: a focus on the role of aberrant stress responses. J. Psychosom. Res., 65, 229–238 (2008).
• 35)  Hughes CG, Patel MB, Pandharipande PP. Pathophysiology of acute brain dysfunction: what’s the cause of all this confusion? Curr. Opin. Crit. Care, 18, 518–526 (2012).
• 36)  Swain DG, O’Brien AG, Nightingale PG. Cognitive assessment in elderly patients admitted to hospital: the relationship between the abbreviated mental test and the mini-mental state examination. Clin. Rehabil., 13, 503–508 (1999).
• 37)  van der Mast RC, Fekkes D. Serotonin and amino acids: partners in delirium pathophysiology? Semin. Clin. Neuropsychiatry, 5, 125–131 (2000).
• 38)  Ali S, Patel M, Jabeen S, Bailey RK, Patel T, Shahid M, Riley WJ, Arain A. Insight into delirium. Innov. Clin. Neurosci., 8, 25–34 (2011).
• 39)  Dantzer R, O’Connor JC, Freund GG, Johnson RW, Kelley KW. From inflammation to sickness and depression: when the immune system subjugates the brain. Nat. Rev. Neurosci., 9, 46–56 (2008).
• 40)  O’Donovan A, Hughes BM, Slavich GM, Lynch L, Cronin MT, O’Farrelly C, Malone KM. Clinical anxiety, cortisol and interleukin-6: evidence for specificity in emotion-biology relationships. Brain Behav. Immun., 24, 1074–1077 (2010).
• 41)  Trzepacz PT. The neuropathogenesis of delirium: a need to focus our research. Psychosomatics, 35, 374–391 (1994).
• 42)  Egberts A, Moreno-Gonzalez R, Alan H, Ziere G, Mattace-Raso FUS. Anticholinergic drug burden and delirium: a systematic review. J. Am. Med. Dir. Assoc., 22, 65–73.e4 (2021).
• 43)  Hshieh TT, Fong TG, Marcantonio ER, Inouye SK. Cholinergic deficiency hypothesis in delirium: a synthesis of current evidence. J. Gerontol. A Biol. Sci. Med. Sci., 63, 764–772 (2008).
• 44)  Gaudreau JD, Gagnon P. Psychotogenic drugs and delirium pathogenesis: the central role of the thalamus. Med. Hypotheses, 64, 471–475 (2005).
• 45)  Alagiakrishnan K, Wiens CA. An approach to drug induced delirium in the elderly. Postgrad. Med. J., 80, 388–393 (2004).
• 46)  Ernst G, Watne LO, Rostrup M, Neerland BE. Delirium in patients with hip fracture is associated with increased heart rate variability. Aging Clin. Exp. Res., 32, 2311–2318 (2020).
• 47)  Echizen M, Satomoto M, Miyajima M, Adachi Y, Matsushima E. Preoperative heart rate variability analysis is as a potential simple and easy measure for predicting perioperative delirium in esophageal surgery. Ann. Med. Surg. (Lond.), 70, 102856 (2021).
• 48)  Chaikijurajai T, Rincon-Choles H, Tang WHW. Natriuretic peptide testing strategies in heart failure: a 2023 update. Adv. Clin. Chem., 118, 155–203 (2024).
• 49)  Miller NE, Rushlow D, Stacey SK. Diagnosis and management of sodium disorders: hyponatremia and hypernatremia. Am. Fam. Physician, 108, 476–486 (2023).
• 50)  Adrogué HJ, Tucker BM, Madias NE. Diagnosis and management of hyponatremia: a review. JAMA, 328, 280–291 (2022).
• 51)  Sterne JA, Sutton AJ, Ioannidis JP, Terrin N, Jones DR, Lau J, Carpenter J, Rücker G, Harbord RM, Schmid CH, Tetzlaff J, Deeks JJ, Peters J, Macaskill P, Schwarzer G, Duval S, Altman DG, Moher D, Higgins JP. Recommendations for examining and interpreting funnel plot asymmetry in meta-analyses of randomised controlled trials. BMJ, 343, d4002 (2011).
• 52)  Japanese Committee for Clinical Laboratory Standards (JCCLS). Clinical Laboratory Standards Guideline, JCCLS, Tokyo (2020).
• 53)  Clinical and Laboratory Standards Institute (CLSI). Fundamentals of Clinical Laboratory Standards: Approved Guidelines (CLSI GP1-A6). CLSI, Wayne, PA (2021).
• 54)  Ministry of Health. Spain. “Good Laboratory Practice Guidelines for Clinical Laboratories. (Guía de buenas prácticas para laboratorios clínicos). Madrid: Gobierno de España. 2020.”: <https://www.sanidad.gob.es.>, accessed 8 April, 2025.