Between 30% and 80% of survivors of critical illness experience cognitive impairment, but the underlying mechanisms remain unknown.
To determine whether intensive care unit (ICU) delirium biomarkers align with the National Institute on Aging–Alzheimer’s Association (NIA-AA) research framework for diagnostic biomarkers for Alzheimer disease and other related dementias (ADRD).
Ovid MEDLINE, PsycInfo, Embase, and the Cochrane Library were systematically searched for articles published between January 1, 2000, and February 20, 2020, on the relationship between delirium and biomarkers listed in the NIA-AA framework. Only studies that addressed delirium in the ICU setting and fluid biomarkers were included in these analyses.
Of 61 256 records screened, 38 studies met inclusion criteria, 8 of which were suitable for meta-analysis. In pooled analysis, significant associations were found between ICU delirium and amyloid β-peptide 1-40 (standard mean difference [SMD], 0.42; 95% CI, 0.09-0.75), interleukin (IL)-1 receptor antagonist (SMD, 0.58; 95% CI, 0.21-0.94), and IL-6 (SMD, 0.31; 95% CI, 0.06-0.56). No significant association was observed in pooled analyses between ICU delirium and the other biomarkers. Few studies have examined ICU delirium and pathologic tau or neurodegeneration biomarkers.
Inflammatory biomarkers and amyloid β are associated with ICU delirium and point to potential overlapping mechanisms between delirium and ADRD. Critical care providers should consider integrating diagnostic approaches used in ADRD in their assessment of post–ICU cognitive dysfunction.
Notice to CE enrollees:
This article has been designated for CE contact hour(s). The evaluation demonstrates your knowledge of the following objectives:
Describe the rationale for examining the overlap in biomarker profile of intensive care unit (ICU) delirium and dementia.
Identify fluid biomarkers associated with ICU delirium and dementia.
Identify areas for future biomarker research in the field of ICU delirium and post-ICU cognitive impairment.
To complete the evaluation for CE contact hour(s) for this article #A21604, visit www.ajcconline.org and click the “CE Articles” button. No CE evaluation fee for AACN members. This expires on July 1, 2023.
The American Association of Critical-Care Nurses is accredited as a provider of nursing continuing professional development by the American Nurses Credentialing Center’s Commission on Accreditation, ANCC Provider Number 0012. AACN has been approved as a provider of continuing education in nursing by the California Board of Registered Nursing (CA BRN), CA Provider Number CEP1036, for 1.0 contact hour.
Between 30% and 80% of survivors of critical illness acquire post–intensive care unit (ICU) cognitive impairment.1-3 The mechanisms underlying this association are still uncertain, but growing evidence suggests that ICU delirium is a major risk factor for post-ICU cognitive impairment.4,5 A crucial step in understanding this relationship is to examine how the biomarker profile of ICU delirium compares with the biomarker profiles of well-characterized dementia subtypes (eg, Alzheimer disease).
The National Institute on Aging and the Alzheimer’s Association have recently proposed a research framework for biomarkers related to Alzheimer disease and other related dementias (ADRD).6 The framework groups biomarkers into those of β-amyloid deposition (A), pathologic tau (T), neurodegeneration (N), and other (X; eg, markers of inflammation, vascular changes), known collectively as AT(N)-X. To examine whether ICU delirium has a biomarker profile comparable to that of ADRD, we conducted a meta-analysis of ICU delirium biomarker studies to determine their alignment with disease state biomarkers of the AT(N)-X framework.
Search Criteria and Strategy
This systematic review and meta-analysis was conducted in accordance with the Preferred Reporting Items for Reviews and Meta-Analyses (PRISMA) guidelines. The search strategy and full inclusion/exclusion criteria were previously described in detail.7 Briefly, Ovid MEDLINE, PsycInfo, Embase, and the Cochrane Library were searched for articles published between January 1, 2000, and February 20, 2020, using a combination of controlled vocabulary, keywords, and Medical Subject Headings terms pertaining to the concepts of delirium and biomarkers in the AT(N)-X framework.6 Inclusion criteria for the search were as follows: (1) patients’ age 18 years or older, (2) use of standardized delirium screening tools or diagnostic assessments, and (3) biomarker measurements listed in AT(N)-X categories (regardless of the modality used). Study exclusion criteria, in brief, were as follows: (1) primary focus on delirium in the following contexts: (a) premorbid or comorbid ADRD, (b) psychiatric disorders, (c) central nervous system disorders other than ADRD, (d) terminal illness, (e) persistent delirium, (f) long-term care settings, and (g) risk of delirium; and (2) non-English-language articles. The aim of excluding studies focused on delirium in the above-listed contexts was to eliminate potentially confounding factors. Persistent delirium, terminal delirium, and delirium in central nervous system disorders other than ADRD may have a different etiology than delirium in the general hospital setting, and individuals in the long-term care setting and with premorbid or comorbid ADRD most likely have abnormal AT(N)-X biomarkers at baseline.
Study selection was conducted by 2 independent, blinded reviewers, and discrepancies were evaluated by consensus panel. Citations were first reviewed for selection criteria based on title and abstract, followed by full-text review. If multiple articles were published from the same cohort, we included data from the study with the most detailed report of biomarker quantities.
Data Extraction and Quality Assessment
Data extraction was conducted by independent, blinded reviewers. Discrepancies were resolved by consensus. Only studies that examined delirium in the ICU setting and fluid biomarkers were included in these analyses. For all studies suitable for meta-analysis, we extracted the means and errors (SDs, SEs, or CIs) and sample sizes of each biomarker for delirium cases and nondelirium controls. If SD was not reported, it was calculated from SEs or CIs. For studies that reported the median, the minimum and maximum values, and/or the first and third quartiles, we estimated the sample mean and SD from these quantities. When a study presented estimates in the figures without providing numerical values, data were obtained from graphs using ImageJ software (http://imagej.nih.gov).
Standardized mean differences (SMDs) and 95% CIs were calculated using DerSimonian and Laird random-effects models. SMDs were chosen as the main outcome because different assays and units were used for biomarkers. Between-study heterogeneity was assessed by a Q statistic and I2 index. The Q statistic was calculated using a χ2 test to quantify the heterogeneity among combined results. Inconsistency was calculated using the I2 index. The I2 values of 25%, 50%, and 75% indicate low, medium, and high heterogeneity, respectively. All statistical significance was defined as 2-sided (α < .05). All statistical analyses were performed with Stata statistical software, version 12 (StataCorp). Significance was set at P less than .05.
In pooled analysis, significant associations were found between ICU delirium and amyloid β-peptide 1-40, interleukin (IL)-1 receptor antagonist, and IL-6.
A detailed flow diagram for inclusion and exclusion of studies per PRISMA guidelines is shown in the Supplemental Figure (available online only at www.ajcconline.org). Methods and results of study quality assessment have been previously published.7
Of the 113 studies that met the criteria for initial systematic review, 38 studies examined the relationship between ICU delirium and fluid biomarkers. All studies focused on serum biomarkers. Nine included both medical and surgical patients, 11 included only medical patients, and 18 included only surgical patients. The Confusion Assessment Method for the Intensive Care Unit (25 studies) and Confusion Assessment Method (8 studies) were the most common methods of delirium assessment. No study assessed the association between “T” (pathologic tau) biomarkers and ICU delirium. Three studies included “A” (β-amyloid [Aβ] deposition) biomarkers. Seven studies examined “N” (neurodegeneration) biomarkers, including neuron-specific enolase (NSE) (4), total tau protein (T-tau) (2), and neurofilament light (1). Thirty-seven studies included “X” biomarkers, and all of those studies examined inflammatory biomarkers.
Biomarkers that met criteria for meta-analysis included serum Aβ1-40, Aβ1-42, T-tau, NSE, C-reactive protein (CRP), interleukin (IL)-1β, interleukin-1 receptor antagonist (IL-RA), IL-6, IL-8, IL-10, IL-17, S-100 calcium-binding protein B (S100β), tumor necrosis factor α (TNF-α), and monocyte chemoattractant protein-1 (MCP-1). The Table and the Figure show meta-analysis of biomarkers studied in association with ICU delirium. A significant association was found between ICU delirium and Aβ1-40 (SMD, 0.42; 95% CI, 0.09-0.75), IL-RA (SMD, 0.58; 95% CI, 0.21-0.94), and IL-6 (SMD, 0.31; 95% CI, 0.06- 0.56). The mean ages of participants in studies included in meta-analyses are shown in the Supplemental Table (available online only).
To our knowledge, this is the first study to conduct a meta-analysis of AT(N) and inflammatory biomarkers in ICU delirium. We found a modest relationship between ICU delirium and elevated Aβ1-40, IL-RA, and IL-6, consistent with previously reported patterns of serum amyloid and inflammatory biomarkers in ADRD.16,17 Few studies examined neurodegenerative biomarkers in the context of ICU delirium, and no studies focused on pathologic tau were found.
Our findings suggest that there may be overlap in the biomarker profile of ICU delirium and ADRD and support prior observations that delirium is associated with persistent cognitive dysfunction in survivors of critical illness,5 including a large observational case-control study (n = 21 000).4 It remains unclear whether delirium (1) accelerates cognitive decline in at-risk individuals with premorbid ADRD pathology or (2) causes de novo ADRD pathology. Further studies directly examining post-ICU cognitive dysfunction and ADRD biomarkers are needed. In contrast to a recent systematic review18 indicating that in studies of critically ill patients, S100β and CRP were associated with delirium, we did not find a significant association between ICU delirium and S100β or CRP in our pooled analyses. This discrepancy may be explained by differences in included studies. As an example, for S100β, only 1 of the 6 studies by Michels et al18 was included in our meta-analysis, whereas the other 2 studies in our meta-analysis were not reported in their systematic review. Similarly, for CRP only 1 study was included in both Michels et al and our meta-analysis. These differences may be explained by differences in search strategy, including our exclusion of non-English-language articles and inclusion of studies that did not use the Confusion Assessment Method or Confusion Assessment Method for the Intensive Care Unit to assess delirium, as well as limiting our meta-analysis to studies that were methodologically comparable.
Although the number of studies was limited for the “T” and “N” categories, we found a significant relationship between Aβ1-40 and ICU delirium. It should be noted, however, that although we excluded studies on premorbid or comorbid ADRD (a known risk factor for delirium), one of the studies included in the meta-analysis of Aβ1-40 and Aβ1-42 contained a delirium group that was older than the nondelirious controls (mean [SD] age, 73 [9.9] vs 67  years; P = .03). This difference could have introduced bias, as amyloid is known to increase with age, even in individuals with normal cognition.19 Although Aβ1-42 is more specific to the pathophysiology of Alzheimer disease,20 Aβ1-40 is the most abundant form of Aβ in the brain, and the concentration ratio of Aβ1-42/Aβ1-40 has also been proposed to be a superior biomarker of Alzheimer disease compared with Aβ1-42 alone.21 Aβ has been associated with increased proinflammatory cytokines in Alzheimer disease,22 although to our knowledge this relationship in the context of delirium has not been investigated. Further study is needed to investigate whether this relationship has implications for the elevated risk of cognitive decline following critical illness.
These findings should be interpreted in the context of their limitations. As this was a meta-analysis of observational studies, the influence of confounding variables cannot be ruled out or fully assessed. Few studies met the criteria for meta-analysis, resulting in only 2 studies being included in the meta-analysis for 7 biomarkers. Heterogeneity was high in the meta-analysis of several biomarkers, including CRP and IL-8.
In summary, Aβ1-40, IL-RA, and IL-6 are associated with ICU delirium and point to potential overlapping mechanisms between delirium and ADRD. Critical care researchers and clinicians should consider integrating diagnostic approaches used in ADRD in their assessment of post-ICU cognitive dysfunction. Additional studies systematically examining the relationship between ICU delirium and biomarkers in the AT(N) framework are needed to provide further insight into the underlying mechanisms of ICU delirium and its relationship with post–intensive care cognitive decline and ADRD.
Drs Chan and Song contributed equally to this work.
Dr Wang is supported by grants from the National Institute on Aging (2P30AG010133 and K23AG062555-01). Dr Lindroth is supported by a grant from the National Heart, Lung, and Blood Institute (5T32HL091816-07). Dr Khan is supported by grants from the National Heart, Lung, and Blood Institute (R01HL131730) and the National Institute on Aging (R01AG055391).
To purchase electronic or print reprints, contact American Association of Critical-Care Nurses, 27071 Aliso Creek Road, Aliso Viejo, CA 92656. Phone, (800) 899-1712 or (949) 362-2050 (ext 532); fax, (949) 362-2049; email, firstname.lastname@example.org.