The association between C-reactive protein and

0 downloads 0 Views 282KB Size Report
International Psychogeriatric Association 2014 doi:10.1017/S1041610213002433. The association between C-reactive protein and delirium in 710 acute elderly ...
C International Psychogeriatric Association 2014 International Psychogeriatrics (2014), 26:5, 717–724  doi:10.1017/S1041610213002433

The association between C-reactive protein and delirium in 710 acute elderly hospital admissions ...........................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................

C.W. Ritchie,1 T.H. Newman,1 B. Leurent2 and E.L. Sampson2,3 1

Centre for Mental Health, Imperial College London, London, UK Marie Curie Palliative Care Research Unit, UCL Mental Health Sciences Unit, University College Medical School, Charles Bell House, London, UK 3 Barnet Enfield and Haringey Mental Health Trust, St Ann’s Hospital, St Ann’s Road, London, UK 2

ABSTRACT

Background: Delirium is a common neuropsychiatric syndrome associated with poor outcomes. Evidence supports a neuroinflammatory etiology, but the role of the inflammatory marker C-reactive protein (C-RP) remains unclear. We investigated the relationship between C-RP and delirium and its severity as well as interaction with medical diagnosis. Methods: From an existing database (710 patients over 70 years old admitted to a Medical Acute Admissions Unit) we analyzed data which included C-RP levels, delirium (using the Confusion Assessment Method), and other clinical and demographic factors. Primary diagnoses were grouped (cardiovascular, musculoskeletal, infection, metabolic, and other). Results: There was a strong association between elevated C-RP and delirium (t = 5.09; p < 0.001), independent of other potential risk factors for delirium (odds ratio (OR) = 1.32 (95% CI: 1.10–1.58) p = 0.003). There was no significant association between C-RP and delirium severity, and between C-RP and delirium in the populations with cardiovascular disease, infection upon admission, or from the metabolic group despite an OR of 2.24 (95% CI: 0.92–5.45). There was an association in the musculoskeletal group (OR 2.19 (95% CI: 1.19–4.02)). Conclusions: There is an association between elevated C-RP and delirium. This is strongest in patients admitted with musculoskeletal disease but not in others, implying that C-RP is involved in the genesis of delirium in musculoskeletal disease, but that other factors or processes may be more important in those with cardiovascular disease or infection. Key words: delirium, organic mental disorders, biomarkers, psychogeriatrics, neuropsychiatry, inflammation, C-reactive protein

Introduction Delirium is a serious and often undiagnosed neuropsychiatric syndrome commonly affecting elderly hospitalized patients (Siddiqi et al., 2006; Collins et al., 2010). It is characterized by an acute onset and fluctuating course with deficits in attention, arousal state, and perception. The cause frequently originates outside the brain, for example, arising secondary to infection, trauma, or metabolic disturbance (Elie et al., 1998; van Gool et al., 2010). The genesis of delirium is multi-factorial but Creactive protein (C-RP) is known to be elevated, non-specifically in multiple conditions that can Correspondence should be addressed to: Dr Craig W Ritchie, Centre for Mental Health, Imperial College London, Claybrook Centre, 37, Claybrook Road, Hammersmith, London W6 8LN, UK. Phone: +44-0207-430-1630. Email: [email protected]. Received 17 Aug 2013; revision requested 25 Sep 2013; revised version received 22 Nov 2013; accepted 25 Nov 2013. First published online 24 January 2014.

precipitate delirium (Clyne and Olshaker, 1999; Hopstaken et al., 2003; Wright and Frier, 2008; Yeh et al., 2012). This implies that C-RP sits as a downstream factor in the genesis of delirium, but upstream from the theorized cholinergic and dopaminergic synaptic dysregulation that generates clinical symptoms (Hshieh et al., 2008; Sanders, 2011; Cerejeira et al., 2012). C-reactive protein is an acute phase protein with widespread clinical utility as a marker for infection, inflammation, and tissue injury (Pepys and Hirschfield, 2003; Black et al., 2004; Macdonald et al., 2007). There is also evidence to support inflammation in the pathogenesis of delirium in both peripheral and central tissues (Cerejeira et al., 2010). Several actions of C-RP can lead to neuroinflammation such as disruption of the blood brain barrier and increase in both central cytokine and microglial activity (Perry, 2004; Uchikado et al., 2004; Quan and Banks,

718

C. W. Ritchie et al.

2007; Alexander et al., 2008; Kuhlmann et al., 2009; Juma et al., 2011). However, evidence for an association between C-RP and delirium is inconsistent and from small clinical populations (Aldemir et al., 2001; Pisani et al., 2007). Moreover, C-RP may be an intermediary factor only in certain pathways, from different origins culminating in delirium. The aim of this study was to describe the association between C-RP and the incidence and severity of delirium in a large sample of elderly patients admitted to an acute hospital, and to determine if there was any interaction between C-RP and delirium by diagnosis as a proxy for upstream etiologies.

Methods To investigate the hypotheses, an existing database was used. Data were gathered for the study on the prevalence and mortality of dementia in the acute hospital setting (Sampson et al., 2009). Recruitment of participants took place at a North London teaching hospital, which serves a diverse population of almost 1.5 million people. During a six-month period all patients aged over 70 years with unplanned acute admission to the Medical Acute Admissions Unit (MAAU) were considered eligible for inclusion. Patients who were admitted to the ward for less than 48 hours or who did not speak adequate English for evaluation by the Confusion Assessment Method (CAM) were excluded. All patients included in the sample were assessed within 72 hours of admission by trained specialists in old age psychiatry, and this assessment (including the CAM; Thomas et al., 2012) was conducted prior to consulting the medical notes (Sampson et al., 2009). The CAM has good discriminating ability to avoid misclassification of dementia as delirium. It can also be used to assess the severity of delirium with a good inter-observer reliability in general wards and outpatients (Inouye et al., 1990; Wei et al., 2008). This was calculated by summing four symptom scores in the CAM: inattention, disorganized thinking, and altered level of consciousness as absent (0), mild (1), or marked (2), while acute onset and fluctuation was measured as absent (0) or present (1), giving a total out of 7 (Wei et al., 2008). Participants’ age, gender, and place of residence were collected from hospital notes, and results from all routine blood tests were recorded. Primary cause of admission was obtained from the Hospital Episode Statistics using ICD-10 diagnosis; this was recorded independently by the hospital coding department and categorized according to

the Ambulatory Care Sensitive Condition system (Sanderson and Dixon, 2000). The Atropine Risk Score (Rudolph et al., 2008) was calculated from all medications at the time of admission. We then categorized each patient into one of the following five diagnostic clusters based on primary diagnosis and medical admission causes: (1) cardiovascular, (2) musculoskeletal, (3) infection, (4) metabolic, or (5) other. The severity of acute illness was measured using the Acute Physiological and Chronic Health Evaluation (APACHE) II score (Knaus et al., 1985). Chronic comorbidity was measured using the Charlson Comorbidity Index (CCI; Charlson, 1987; de Groot et al., 2003). Data Analysis An independent t-test was used to investigate the relationship between C-RP and delirium. This was followed by logistic regression to adjust for predisposing or causal factors of delirium possibly related to C-RP: age, gender, CCI score, APACHE score, diagnosis at admission, and atropine equivalence of medication prescribed upon admission. Investigation of a possible difference in the association between C-RP levels and delirium between diagnosis categories was undertaken using a logistic regression between C-RP and delirium stratified by diagnostic category. Test for a significant modifying effect (interaction) was performed. The relationship between C-RP and delirium severity was also investigated using an analysis of variance (ANOVA) and the nonparametric Spearman’s rank correlation coefficient. All analysis took place in SPSS 20.0 or STATA version 12 as appropriate. The study was authorized by the hospital’s ethics committee (Sampson et al., 2009).

Results A total of 710 participants were included, of whom 87 (12.3%) were observed to have delirium within 72 hours of admission. The mean age was 83.05 years (SD = 7.4; minimum age = 70 years, maximum age = 101 years), 419 (59%) were females, 619 (87.2%) were described as of White British ethnicity, and 57% were smokers (Table 1). C-RP levels and delirium C-reactive protein levels were highly skewed, so these were log-transformed. There was a very strong association between the presence of delirium and raised C-RP on admission (t = 5.09; p < 0.001; Table 2).

C-reactive protein and delirium

719

Table 1. Sample characteristics WHOLE

PARTICIPANTS

PARTICIPANTS

SAMPLE

WITH DELIRIUM

WITHOUT DELIRIUM

(N = 710)

(N = 87)

(N = 623)

STATISTICS

p-V A L U E

...........................................................................................................................................................................................................................................................................................................................

Age Gender (% Male) Ethnicity (White British) Smoking status Current Ex-smoker Never smoked Apache total score (n = 656) Charlson comorbidity index

83.05 (SD = 7.4) 291/710 (41%) 619/710 (87.2%)

86.7 (SD = 7.26) 33/87 (37.9%) 78/87 (89.6%)

82.5 (SD = 7.29) 258/623 (41.4%) 541/623 (86.8%)

65/684 (9.5%) 296/684 (43.3%) 323/684 (47.2%) 12.1 (SD = 3.7) 2.8 (SD = 2.1)

4/80 (5%)

61/604 (10.1%) 266/604 (44%) 277/604 (45.9%) 11.8 (SD = 3.4) 2.8 (SD = 2.2)

30/80 (37.5%) 46/80 (57.5%) 14.4 (SD = 4.4) 2.8 (SD = 1.5)

t = –4.9