Pre-operative Biomarkers and Imaging Tests as Predictors of Post-operative Delirium in Non-cardiac Surgical Patients

A Systematic Review

Farrah Ayob; Enoch Lam; George Ho; Frances Chung; Hossam El-Beheiry; Jean Wong

BMC Anesthesiol. 2019;19(25)

Abstract and Introduction

Abstract

Background: Post-operative delirium (POD) is a common post-operative complication in elderly individuals and imposes a significant health and financial burden. Identifying predictive biomarkers may help understand the pathophysiology of POD. Our objective is to summarize the evidence of pre-operative biomarkers and imaging tests to predict POD in patients undergoing non-cardiac surgery.

Methods: A systematic search of English language articles in MEDLINE, EMBASE, Cochrane Database, PsychINFO, PubMed and ClinicalTrials. Gov up to January 2018 was performed. Studies that used biomarkers or imaging tests to predict POD and a validated POD assessment tool were included. Animal studies, paediatric, cardiac and intracranial surgery were excluded. Risk of bias was assessed using the Quality In Prognosis Study tool.

Results: Thirty-four prospective cohort studies involving 4424 patients were included. Nineteen studies described serum tests [Interleukin-6, Insulin-like Growth Factor 1, C-Reactive Protein (CRP), cholinesterases, apolipoprotein-E genotype, leptin, hypovitaminosis, hypoalbuminaemia, gamma-amino butyric acid], 10 described cerebral-spinal fluid tests (monoamine precursor, melatonin, acute phase proteins, S100B and neurofibrillary tangles), and 5 described imaging tests. Two studies had high risk of bias due to unclear outcome measurement and study participation. CRP was significantly associated with POD in 5 studies. Other biomarkers were either examined by only a single study or two or more studies with conflicting results.

Conclusion: CRP is the most promising biomarker associated with POD. However, we are still in the early stages in identifying biomarkers and imaging tests that may further understanding of the pathophysiology of POD.

1 2 3 4 5

Next Section

Abstract and Introduction
Background
Methods
Results
Discussion
References

Table 1. Summary of studies: Pre-operative serum tests
Author, Year, Country	Study design	n	Cut-off age	Type of surgery	Included patients with dementia or cognitive impairment	Multivariate regression analysis performed	Pre-operative tests	POD assessment, Days post-surgery assessment	POD- mean age, % female	No POD – mean age, % female	Conclusion	Pre-op Test Predictive of POD
Biomarkers
Capri, 2014, Italy [8]	PC	74 (37 w/POD, 37 w/o POD)	> 65	Any elective and emergency surgery, excluding cardiac	No	Yes	Plasma cytokine concentrations (TNF-a, IL-1b, IL-2, IL-6, IL-8, and IL-10)	CAM and DRS, D1,2,3 and 6	79.2 yr., 45.9%	76.4 yr., 54%	High serum IL-6 (> 9 pg/mL) independently predicts POD. OR 4.9, [1.6–14.63; P < 0.0005].	Yes
Westhoff, 2013, Netherlands [9]	PC	61	> 75	Emergency hip fracture	Yes	MA logistic regression analysis not performed but no significant correlation was found between age, cognitive function and levels of CSF cytokines.	42 CSF and serum cytokines and chemokines	CAM, DRS-R- 98, D1–5	82.9 yr., 69.9%	84.6 yr., 68.4%	Pre-op CSF Fms-like tyrosine kinase-3, IL-1r antagonist and IL-6. Pre-op serum IL-6 was significantly higher in patients with POD. Median levels in POD vs .no POD: 48.13 vs 23.16, P = 0.021.	Yes
Dillon, 2017, Israel and UK [10]	PC	566, Sub cohort from SAGE	> 70, Pooled cohort (combined match pairs)	Elective major non-cardiac	No	Conditional logistic regression and sensitivity analysis performed.	Serum CRP	CAM and validated chart review, POD = D1–2	77.6 yr., 56%	77.2 yr., 56%	Elevated pre-op CRP associated with POD (pre-op median paired difference with controlled subjects of 1.97 mg/L, P = 0.02).	Yes
Neerland, 2016, Norway & UK [11]	PC	151 Oslo 99, Edinburgh 52	> 60	Emergency hip fracture	Yes	No	Serum and CSF CRP, IL-6 and soluble IL-6 receptor	CAM O: D1–5, E: D1–4, 7, 10,14	85 yr., 70%	83 yr., 80%	Significantly high CSF levels of CRP in POD vs. no-POD, median 0.05mcg/mL vs. 0.01mcg/mL	Yes
Vasunilashorn, 2017, USA [12]	PC	560 (sub cohort from SAGE) POD = 25% of participants	> 70	Major non-cardiac surgery	No	No	Serum CRP	CAM, D1 until discharge	77.7 + − 5.0 yr., 60%	76.5 + − 5.2 yr., 58%	Subjects with pre-op CRP of ≥3 mg/L had a 1.5 greater risk of delirium than subjects with CRP; 4 more delirium days (P < 0.001) and 1.4 times greater risk of prolonged LOS.	Yes
Xiang, 2017, China [13]	PC	160 POD = 39(24.4%)	> 65	Laparoscopic surgery for colon carcinoma	No	Yes	Serum CRP	CAM-ICU D1–3 and 7	72.2 + − 5.8 yr., 41%	69.4 + − 7.1 yr., 39.7%	Pre-op CRP level: independent predictor of POD. POD vs. no POD CRP levels; 3.8 vs. 2.4 mg/L (OR: 5.87; 95% CI 2.22–11.4, P = 0.018).	Yes
Bohner, 2003, Germany [38]	PC	153	N/S	Elective Vascular surgery	No	No. Only Univariate analysis performed.	Pre-op: CRP, White cell count, platelets, LFTs, Creatinine, Urea, coagulation, Intra-op: BP, BG, glucose	DSM IV criteria, DRS > 12, D1–7	63.7 + − 10.3 yr.,	68.3 + − 8.5 yr.,	Pre-op CRP significantly higher in POD vs. no POD; 3.4 vs. 1.7 mg/dL, P = 0.03. AT III (98% vs. 106%, p = 0.02), lower Hb 13.7 vs. 14.3 g/dL, P = 0.02).	Yes
Lemstra, 2008, Netherlands [14]	PC	68	> 70	Elective hip replacement	Yes	No. Patients were matched for age, severe illness and MMSE score < 24 using a statistical analysis.	Serum CRP, IL-6, IGF-1	CAM and DSM-IV daily	80 yr., 55.6%	78.5 yr., 74%	No difference between POD vs. no POD levels: CRP = 5.3 vs. 3 mgL^{− 1}, P = 0.523; IL-6 = 3.6 vs. 3.0 pgmL^{− 1}, P = 0.121; IGF-1 = 14.4 vs. 12.9 nmolL ^{− 1}, P = 0.675.	No
Shen, 2016, China [15]	PC	140	> 65	Elective open GI tumour resection	No	Yes	Serum IGF-1	CAM, DRS-R98. D2,3	73.8 yr., 52.8%	68.8 yr., 58.7%	Significantly low serum IGF-1 (POD vs. no-POD, 50.4 vs. 67 ng/mL), OR 2.52 (1.19–5.43).	Yes. Cut-off level 52.94 ng/ml, sensitivity of 80.8%, specificity of 80.6%.
Yen, 2016, USA [37]	PC	98	> 65	Elective knee replacement	No	Yes	Serum IGF-1	DSM-IV, DRS-R98	72.5 yr., 59%	73.7 yr., 50%	No association. Median levels in POD vs. no POD = 62.6 vs. 65.9 ng/mL, P = 0.141.	No
Cerejeira, 2011, Portugal [16]	PC	101	> 60	Elective total hip replacement	No	No	Plasma AChE and BuChE activity	CAM, D1,2,3	73.7 yr., 60%	72.7 yr., 46%	AChE and BuChE 24 and 32% lower in patients with vs. without POD, respectively. Pre-operative differences between the two groups were controlled.	Yes
Cerejeira, 2012, Portugal [17]	PC	101	> 60	Elective total hip replacement	No	No	Plasma AChE and BuChE and inflammatory mediator levels (CRP, IL-1b, and TNF-a)	CAM, D1,2,3	73.6 yr., 60%	72.7 yr., 45.3%	Low baseline plasma cholinesterase activity associated with POD and positively correlates with high CRP, IL-6 and Pro-inflammatory/Anti-inflammatory ratio.	Yes
Chen, 2014, China [18]	PC	186	> 65	Emergency hip fracture	No	Yes	Plasma leptin	CAM, D1,2,3,7 and 1 month	80.1 yr., 72%	74.7 yr., 77%	Plasma leptin level is significantly lower in POD vs. no-POD, 4 vs. 7.5 ngmL^{− 1}, P < 0.001). Hypoleptinaemia is an independent predictor of POD [OR 0.385, CI 0.28–0.517].	Yes. Sensitivity 72.2%, Specificity 91.7%
Cunningham, 2017, UK [19]	PC	315	> 65	Elective hip or knee replacement	No	Yes	Serum ApoE4 allele carriage and neuropsychological tests	CAM, D1,2,3	76.9 yr., 65%	74 yr., 27.1%	ApoE4 genotype is not associated with POD.	No
Leung, 2007, USA [20]	PC	190	> 65	Elective major non-cardiac	No	Yes	Serum apolipoprotein (APOE) genotype	CAM, D1–2	74.2 yr., NR	72.3 yr., NR	Presence of at least one copy of APOE e4 allele is associated with increased risk of POD that persists 2 days post-op. (OR 3.64, 95% CI 1.51–8.77).	Yes
Vasunilashorn, 2015, USA [21]	PC	557	> 70	Any Elective major non-cardiac	No	Sensitivity analyses performed to test if dementia has any influence on association between ApoE and POD.	Serum Apolipoprotein E: e2, e4 carrier vs non-carrier and three category ApoE genotypes (e3e3, e3e4, e4e4, e2e2, e3e2)	CAM and validated chart review daily			ApoE genotype has no association with incidence, severity or duration of POD, RR for E4 = 1.0, CI 0.7–1.	No
Torbergsen, 2015, Norway & Scotland [22]	PC	115	NR	Emergency hip fracture	No	Yes	Serum vitamin levels (Vitamin A, B1, B6, vitamin B12, Folic acid, vitamin C, D, K)	CAM, D1–5	84.8 yr., 71%	80.6 yr., 79%	Vitamin D deficiency (< 50 nmol/L) was independent predictor of POD. (Mean Vitamin D in POD vs. no-POD, 41 vs. 52 nmolL⁻¹, P = 0.05).	Yes
Scholtens, 2017, Netherlands [23]	PC	144	> 65	Emergency hip fracture	No	No	Morning plasma melatonin	CAM	85.5 yr., 75%	82.5 yr., 69%	Morning melatonin not associated with POD (p = 0.35)	No
Scholtens, 2016, Netherlands [7]	PC	60	> 65	Emergency hip fracture	Yes	No but confounders (age and cognitive impairment) analysed and did not show any difference in CSF melatonin levels.	CSF melatonin	Delirium Observation Screening Scale, and DSM-IV, NS	86.4 yr., 83.3%	83.4 yr., 66%	Pre-operative CSF melatonin did not differ between POD and no-POD groups. No MA performed	No
Wyrobek, 2017, USA [24]	PC	77	> 70	Major Spine surgery Median age = 75 years, 47% were female	No	Yes	Intra-op serum brain derived neurotrophic factor (BDNF)	CAM, D1–4 CAM-ICU Validated chart review	NR	NR	Subjects with POD vs no. without POD had a greater percentage of BDNF decline from baseline, median 75% (IQR 51–82) vs. 50% (IQR 14–79, P = 0.03).	Yes
Zhang, 2017, China [25]	PC	700 111 (15.9%)	> 65	Elective non -cardiac surgery who were admitted to ICU	No	Yes	Pre-op serum albumin levels	CAM-ICU D1–7	76.2 + −7.8 yr., 47.7%	74.0 + − 6.6 yr., 38%	Pre-op severe hypoalbuminemia (< 30.0 g/L) was associated with increased risk of POD (OR 2.727, 95% CI 1.28–5.797, P = 0.009) .	Yes
Hall, 2016, Norway and Scotland [26]	PC	139 (Oslo- 85, Edinburgh- 54)	Oslo- > 60, Edin > 61	Emergency hip fracture	Yes	Yes	Serum and CSF neopterin.	CAM, D1–5	85 yr., 70%	82 yr., 77%	Higher pre-op CSF neopterin in POD vs. no-POD. (Median 29.6 vs 24.7 nmolL^{− 1}, P = 0.003).	Yes
Hov, 2016, Norway [27]	PC	120	NR	Emergency hip fracture	Yes	No	Serum and CSF albumin. q-albumin (ratio of CSF albumin to serum albumin)	CAM, D1–5	85 yr., 70%	83 yr., 83%	CSF barrier dysfunction (q-albumin> 10.2). Significant difference in POD vs. no POD group, n = 11 (16% vs 0%, P = 0.022).	Yes
Watne, 2014, Norway [28]	PC	151 Oslo = 99 Edin = 52	Oslo- NR, Edin > 60, 84.	Emergency hip fracture	No	Yes	Serum and CSF anticholinergic activity	CAM D1–5	Oslo: 75% Edin: 60%	Oslo 72%, Edin: 75%	Serum or CSF anticholinergic activity is not an important mechanism in POD.	No
Watne, 2016, Norway [29]	PC	77	> 70	Emergency hip fracture	No	Yes	Serum and CSF monoamine precursors	CAM D1–5	86 yr., 70%	84 yr., 61%	Higher CSF monoamine precursors (tryptophan, tyrosine, phenylalanine, methionine, 5-HIAA) in delirium patients suggests high monoaminergic activity in CNS during delirium.	Yes
Hov 2017, Norway [30]	PC	Hip fracture = 98 Elective surgery = 50	> 70	Hip fracture and elective surgery (gynae, urology or orthopaedic	No	No	Serum and CSF S100B and phosphorylated tau (P-tau) concentration	CAM D1–5	Prevalent delirium- 85 yr., 75%, Incident delirium- 88 yr., 56%	84 yr., 78%	Significant difference in CSF S100B in patients with vs. without incident delirium (1.38 vs. 1.08 μgrams/L, P = 0.013).	Yes
Hall, 2013, Scotland [31]	PC	45	> 60	Emergency hip fracture	Yes	No	Serum and CSF S100B	CAM and DRS-R98, D1–4,7,10–14	81.3 yr., 63%	78.9 yr., 73%	No significant difference in log10 CSF S100B in POD vs no-POD during the 2-week period (mean: −0.239 vs −0.308 respectively; student's t-test t = 1.25, df = 43, P = 0.218).	No
Westhoff, 2015, Netherlands [32]	PC	53	> 75, mean age 83.1, 67.9%	Emergency hip fracture	Yes	Yes	CSF proteins (low complement factor C3, fibulin-1, 1-beta-1,3 N-acetylglucosaminyl-transferase in POD, high neural cell adhesion molecule-2, fibrinogen, zinc-α-2-glycoproteinandhaptoglobin level in POD)	CAM, D1–5	NR	NR	Discrepencies of results between proteomic ccohort and validation cohort.	No
Witlox, 2011, Netherlands [33]	PC	76	> 75	Emergency hip fracture	Yes	Yes	CSF levels of senile plaques B amyloid and neurofibrillary tangles (Ptau)	CAM, D1–5	84.7 yr., 67%	82.4 yr., 67%	CSF markers for plaque and tangle formation are not strongly associated with POD.	No

PC Prospective cohort, n number, POD Post-operative delirium, TNF tumour necrosis factor, IL interleukin, CAM confusion assessment method, DRS delirium rating scale, D1 day 1, AChE acetylcholinesterase, BChE butylcholinesterase, CRP c-reactive protein, MMSE mini mental state examination, SAGE Successful Ageing after Elective Surgery, NR not recorded, AT III angiotensin III, Hb hemoglobin, 5-HIAA 5-hydroxyindoleacetic acid, GI gastrointestinal, CSF cerebral spinal fluid, CAM-S Confusion Assessment Method- Short Form, DSM-IV Diagnostic and Statistical Manual 4th, IGF Insulin Like Growth Factor, PACU post-anaesthesia care unit, GA general anaesthesia, ICU intensive care unit, GABA gamma amino butyric acid, Fms Fat Mobilising substance, OR odds ratio, CBF cerebral blood flow, MA multivariate analysis

Table 2. Summary of studies: imaging tests
Author, Year, Country	Study design	n	Cut-off age	Type of surgery	Included patients with dementia or cognitive impairment	Multivariate regression analysis performed	Pre-operative tests	POD assessment, Days post-surgery assessment	POD- mean age, % female	No POD – mean age, % female	Conclusion	Pre-op Test Predictive of POD
Imaging
Cavallari, 2016, USA [ 34]	PC	136	> 70,	Elective non-cardiac	No	No but no significant difference were found between subjects (age, gender, vascular comorbidities)	MRI DTI of brain	CAM and validated chart review, daily	NR, 69%	NR, 56%	Significant association of DTI abnormalities in a variety of brain regions with POD. Mean differences in DTI indices between POD and non-POD = 3–8%, P < 0.05.	Yes
Hshieh, 2017, USA [ 35]	PC	146	> 70	Elective major non-cardiac	No	No	MRI of global and regional cerebral blood flow	CAM and validated chart review	NR, 69%	NR, 56%	No significant association between MRI of cerebral blood flow with POD in both adjusted and unadjusted analyses.	No
Miller, 2017, USA [ 40]	PC	142	> 65	Elective general surgery	No	Yes	CT of total psoas area	Retrospective medical record review	NR	NR	Low total psoas area is associated with POD (OR 3.12, 95% CI = 1.02–9.56, P = 0.046).	Yes
Root 2013, USA [ 41]	RC	48	NR	Anatomic lung resection for non-small cell carcinoma	No	No	MRI – WMHB and CA	Doctors, psychiatry and nurses detect confusion, disorientation, hyperactivity. D1–4	73.39 yrs., 56.5%	73.63 yrs., 54%	WMHB higher in POD vs no POD group (mean 0.01% vs 0.005% of total intracranial volume) p = 0.017.	Yes
Racine 2017, USA [ 36]	PC	145 (subset of SAGES)	> 70	Elective major non-cardiac	No	Yes	MRI of cortical thickness in Alzheimer's Disease signature regions	CAM, validated chart review	77 +/− 4.3 yrs., 69%	76 +/− 4.6 yrs., 58%	Alzheimers' Disease signature cortical thickness did not predict POD (OR = 1.15, 95% CI [0.8, 1.6].	No

PC prospective cohort, RC retrospective cohort, n number, POD post-operative delirium, NR not recorded, CT computed tomography, OR odds ratio, CI confidence interval, p probability, RC randomized controlled, MRI magnetic resonance imaging, DTI diffusion tensor imaging, CAM Confusion Assessment method, D1–4 Day 1–4, SAGES Successful Ageing After Elective Surgery, yrs. years, WMHB white matter hyper-intensity burden, CA cerebral atrophy, MA Multivariate analysis

Table 3. Summary of risk of bias assessment
Domains	Low risk n (%)	Medium risk	High risk
Study participation	27 (79.4%)	6 (17.6%)	1 (2.9%)
Study attrition	29 (85.3%)	5 (14.7%)	0 (0)
Prognosis factor measurement	34 (100%)	0 (0)	0 (0)
Outcome measurement	33 (97.1%)	0 (0)	1 (2.9%)
Confounding measurement	26 (76.5%)	8 (23.5%)	0 (0)
Statistical analysis	34 (100%)	0 (0)	0 (0)

Table 4. Summary of association of biomarkers and imaging tests with post-operative delirium
Biomarkers	Number of studies with positive association with post-operative delirium	References	Number of studies without an association with post-operative delirium	References
IL-6	2	6, 7	1	12
CRP	5	8, 9, 10, 11,,34	1	12
IGF-1	1	13	2	12,35
Angiontensin III	1	34
Hemoglobin	1	34
Reduced Cholinesterase activity	2	14, 15	0
Brain derived neurotrophic factor (BDNF)	1	22	0
Anticholinergic activity	0		1	26
Serum leptin	1	16
APO E 4 genotype	1	18	2	17, 19
Hypovitaminsosis (Vitamin D deficiency)	1	20	0
Melatonin	0		2	5, 21
Hypoalbuminemia	1	23	0
Neopterin	1	24	0
Monoamine precursor	1	27	0
CSF Acute phase proteins	0		1	30
S100B	1	28	1	29
Plaques/tangles	0		1	31
CSF barrier dysfunction	1	25	0
Magnetic Resonance Imaging	2	32, 40	2	33, 41
Computed Tomography Imaging	1	38	0

IL-6 interleukin-6, CRP C-reactive protein, IGF-1 insulin growth factor-1, APO-E alolipoprotein-E, GABA gamma aminobutyric acid

Supplementary Table. QUIPS Risk of bias assessment.
Study	Study Participation	Study Attrition	Prognostic Factor Measurement	Outcome Measurement	Confounding measurement	Statistical Analysis
Lemstra et al. 2008 (12)	L	L	L	L	M	L
Cerejeira et al. 2011 (14)	L	L	L	L	L	L
Witlox et al. 2011 (31)	L	M	L	L	M	L
Cerejeira et al. 2012 (15)	M	L	L	L	L	L
Leung et al. 2007 (18)	L	L	L	L	L	L
Westhoff et al. 2013 (7)	M	L	L	L	M	L
Capri et al. 2014 (6)	L	L	L	L	L	L
Chen et al. 2014 (16)	L	M	L	L	L	L
Watne et al. 2014 (26)	L	L	L	L	L	L
Torbergsen et al. 2015 (20)	L	L	L	L	M	L
Vasunilashorn et al. 2015 (19)	L	L	L	L	L	L
Westhoff et al. 2015 (30)	M	M	L	L	L	L
Cavallari et al. 2016 (32)	L	L	L	L	L	L
Hall et al. 2016 (24)	L	L	L	L	L	L
Hov et al. 2016 (25)	L	L	L	L	L	L
Neerland et al. 2016 (9)	L	L	L	L	M	L
Scholtens et al. 2016 (5)	L	L	L	L	L	L
Shen et al. 2016 (13)	L	L	L	L	L	L
Watne et al. 2016 (27)	L	L	L	L	M	L
Yen et al. 2016 (35)	L	L	L	L	M	L
Cunningham et al. 2017 (17)	L	L	L	L	L	L
Dillon et al. 2017 (8)	L	L	L	L	L	L
Hshieh et al. 2017 (33)	L	L	L	L	L	L
Miller et al. 2017 (36)	H	L	L	H	L	L
Hall et al. 2013 (29)	L	M	L	L	M	L
Hov et al. 2017 (28)	L	L	L	L	L	L
Scholtens et al. 2017 (21)	M	L	L	L	L	L
Vasunilashorn et al. 2017 (10)	L	L	L	L	L	L
Wyrobek et al. 2017 (22)	L	L	L	L	L	L
Xiang et al. 2017 (11)	L	L	L	L	L	L
Zhang et al. 2017 (23)	L	L	L	L	L	L
Bohner et al. 2003 (34)	M	M	L	L	L	L
Racine et al. 2017 (38)	L	L	L	L	L	L
Root et al. 2013 (37)	M	L	L	L	L	L

H, high; M, medium; L, low bias

processing....

Pre-operative Biomarkers and Imaging Tests as Predictors of Post-operative Delirium in Non-cardiac Surgical Patients

A Systematic Review

Abstract and Introduction

Abstract

Tables

Tables

Tables

Tables

Tables

References

Authors and Disclosures

Authors and Disclosures

Email This

Feedback