Phenomenological and biological correlates of improved cognitive function in hospitalized elderly medical inpatients

Archives of Gerontology and Geriatrics 59 (2014) 593–598

Contents lists available at ScienceDirect

Archives of Gerontology and Geriatrics journal homepage: www.elsevier.com/locate/archger

Phenomenological and biological correlates of improved cognitive function in hospitalized elderly medical inpatients Dimitrios Adamis a,b,f,*, David Meagher c,1, Adrian Treloar d,e,2, Colum Dunne c,3, Michael Larvin c,4,5, Finbarr C. Martin a,6, Alastair J.D. Macdonald d,7 a

Department of Ageing and Health, Guy’s and St Thomas’ NHS Foundation Trust, London, UK Sligo Mental Health Services, Clarion Road, Sligo, Ireland Cognitive Impairment Research Group, Centre for Interventions in Infection, Inflammation & Immunity (4i), Graduate-Entry Medical School, University of Limerick, Ireland d Institute of Psychiatry, King’s College, London, UK e Department of Old Age Psychiatry, Oxleas NHS Trust, London, UK f Research and Academic Institute of Athens, Greece b c

A R T I C L E I N F O

A B S T R A C T

Article history: Received 5 October 2013 Received in revised form 11 July 2014 Accepted 11 August 2014 Available online 19 August 2014

Deterioration of cognitive ability is a recognized outcome following acute illness in older patients. Levels of circulating cytokines and APOE genotype have both been linked with acute illness-related cognitive decline. In this observational longitudinal study, consecutive admissions to an elderly medical unit of patients aged 70 years were assessed within 3 days and re-assessed twice weekly with a range of scales assessing cognitive function, functional status and illness severity. Cytokines and APOE genotype were measured in a subsample. Improvement was defined as either a 20% or three points increase in mini mental state examination (MMSE). From the 142 participants 55 (39%) experienced cognitive improvement, of which 30 (54.5%) had delirium while 25 had non-delirious acute cognitive disorder. Using bivariate statistics, subjects with more severe acute illness, lower insulin-like growth factor-I (IGF-I) levels and more severe delirium were more likely to experience a 20% improvement in MMSE scores. When the criterion of cognitive improvement was a 3 point improvement in MMSE, those with more severe delirium, females and older were more likely to be improved. Longitudinal analysis using any criterion of improvement indicated that improvement was significantly (p < .05) predicted by higher levels of IGF-I, lower levels of IL-1 (alpha and beta), lack of APOE epsilon 4 allele, and female gender. In conclusion, cognitive recovery during admission is not exclusively linked to delirium status, but reflects a range of factors. The character and relevance of non-delirious acute cognitive disorder warrants further study. ß 2014 Elsevier Ireland Ltd. All rights reserved.

Keywords: Cognition Delirium Elderly APOE Cytokines

* Corresponding author at: Sligo Mental Health Services, Clarion Road, Sligo, Ireland. Tel.: +353 719144829; fax: +353 719144177. E-mail addresses: [email protected] (D. Adamis), [email protected] (A. Treloar), [email protected] (C. Dunne), [email protected] (F.C. Martin), [email protected] (Alastair J.D. Macdonald). 1 Teaching and Research in Psychiatry, Cognitive Impairment Research Group, Centre for Interventions in Infection, Inflammation & Immunity (4i), Room 01-017 Graduate-Entry Medical School, University of Limerick, Ireland. 2 Old Age Psychiatry, Oxleas NHS Trust, Room 19, Memorial Hospital, Shooters Hill, London SE18 3RZ, UK. Tel.: +44 020 8836 8520; fax: +44 020 8856 5359. 3 Tel.: +353 061 234703; fax: +353 061 233778. 4 Graduate Entry Medical School (Room GEMS 3-021), University of Limerick, Limerick, Ireland. 5 University Hospitals of Limerick Group, Dooradoyle, Limerick, Ireland. 6 Guy’s and St Thomas’ NHS Foundation Trust, Elderly Care Unit, St Thomas’ Hospital, London SE1 7EH, UK. Tel.: +44 020 7188 2515. 7 Old Age Psychiatry, Institute of Psychiatry, King’s College, Room HO.09, HSPR Department, David Goldberg Building, de Crespigny Park, London SE5 8AP, UK. Tel.: +44 203 228 7735. http://dx.doi.org/10.1016/j.archger.2014.08.007 0167-4943/ß 2014 Elsevier Ireland Ltd. All rights reserved.

1. Introduction Impaired cognitive function is a common accompaniment of acute physical illness and can reflect the impact of factors such as physical frailty, fatigue, pain, the direct effect of particular morbidities and exposure to medication or surgery (Cole, You, McCusker, Ciampi, & Belzile, 2008; Avidan & Evers, 2011). The classical construct to describe such disturbances is delirium which involves a complex constellation of neuropsychiatric and cognitive disturbances of acute onset and fluctuating course. However, impaired cognitive performance without delirium is also common, and may reflect a ‘normal’ response to the stress of illness (Jones, Griffiths, Slater, Benjamin, & Wilson, 2006; Torgersen, Hole, Kvale, Wentzel-Larsen, & Flaatten, 2011; Wilson & Morley, 2003). Moreover, the character of neuropsychological impairment in these states may differ from delirium (Lowery, Wesnes, Brewster, & Ballard, 2008).

594

D. Adamis et al. / Archives of Gerontology and Geriatrics 59 (2014) 593–598

The traditional concept of delirium emphasises reversibility as a typical element of the syndrome but for many patients, especially elderly, delirium can be associated with persistent cognitive impairment or so-called long term cognitive impairment (LTCI) (MacLullich, Beaglehole, Hall, & Meagher, 2009). Moreover, studies suggest that the reversibility of cognitive impairment relates more closely to the severity of acute illness than to the presence or severity of delirium (Fields, MacKenzie, Charlson, & Perry, 1986; Inouye et al., 2006). There is, therefore, a need to clarify the relationship between biological and other clinical features with changing cognitive function in elderly medical patients without making assumptions as to the role of particular diagnostic concepts (such as delirium). One approach is to define cognitive impairment according to MMSE scores, with studies to date favoring either a three point improvement (Inouye et al., 2006) or 20% increase from the worst previous total MMSE score (Treloar & Macdonald, 1997a, 1997b) as indicative of significant cognitive improvement. In this paper we investigate change in cognitive function using both definitions. Biological factors are also relevant to changing cognitive function. Impaired cognitive function is associated with altered expression of inflammatory markers (Simone & Tan, 2011) but their direct role or function as mere epiphenomena remains unclear, such that understanding the pathobiology of cognition during acute physical illness requires studies which examine a range of biological and other clinical factors, and investigating how they relate temporally to changing cognitive status. Moreover, studies to date have focused upon the role of biological measures in deteriorating cognitive function rather than during the process of cognitive recovery, which requires more detailed investigation. In addition carriers of the APOE epsilon 4 allele have been found to have worse outcomes, higher mortality and reduced functional improvement after head injury, intracerebral hemorrhage, or stroke. Studies of APOE genotype in delirium have found conflicting results but a meta-analysis showed that the possession of the APOE epsilon 4 allele has a small, non significant effect in the presence or absence of delirium (Adamis & Macdonald, 2009). However, genetic factors like APOE genotype may be implicated in cognitive recovery by maturing amyloid deposition (Marx, Blasko, Grubeck-Loebenstein, 1999) resulting in irreversible cerebral damage. Similarly previous research (Adamis, Treloar, Gregson, Macdonald, & Martin, 2011) shows a relationship between APOE genotype, cognition and functional ability in medical inpatients. The aims of this study were: (a) to investigate the relationship between cognitive improvement [defined as either (i) a 20% improvement in MMSE score or (ii) a 3-point improvement] and a range of phenomenological and biological factors (cytokine levels, functional disability, severity of illness, delirium, delirium severity, possession of APOE epsilon 4 genotype, gender and age), and (b) to explore the relationship between delirium status (CAM positive or negative) and changing cognitive function (MMSE scores) during admission.

known terminal illness, clinically severe aphasia, severe hearing or visual impairment, were intubated or did not speak English. 2.1. Clinical assessments Each participant was assessed between day one and day three of their admission: (hereafter called the ‘‘first assessment’’) and, subsequently, on three occasions (assessments two, three, four) separated by three-day intervals before a final (fifth) assessment at 28 days after first assessment, if still in hospital. The maximum number of assessments was five. Each participant was assessed with (a) The Mini Mental State Examination (MMSE), (b) The Confusion Assessment Method (CAM), (c) The Delirium Rating Scale (DRS), (d) The APACHE II and its subscale Acute Physiology Score APS, (e) dementia was diagnosed according to DSM-IV criteria using all available information, including collateral history where available, (f) physical disability was assessed using the Barthel Index scale at the first and last assessment and with the Frailty Scale at first assessment only. 2.2. Biological measures The APOE genotype was analysed in all patients who consented to provide blood samples for laboratory tests (n = 116). Circulating cytokines and insulin like growth factor-1 (IGF-I) were measured in the first 60 subjects but not subsequent patients, due to financial constraints. These included levels of circulating interferon-g (IFNg), interleukin-1 (IL-1) and its receptor, interleukin-6 (IL-6), leukemia inhibitor factor (LIF), tumor necrosis factor a (TNF-a), and IGF-I, were estimated at each of the first four assessments. 2.3. Statistical analyses Data were analysed using SPSS v19 (IBM). Non-parametric tests of significance were used where appropriate, and a p value of <.05 was accepted as statistically significant. Cognitive improvement was defined as either a 20% or 3-point improvement between the MMSE score at the first assessment and the subsequent best score. A score of 25 or 27 or less is necessary to allow for a subsequent 20% or 3 point improvement, respectively. Therefore, for these analyses, those with MMSE scores greater than these at first assessment were excluded. Factors associated with cognitive reversibility by any of the above definitions during hospitalization were identified using a Generalized Estimating Equations (GEE) model with logit link function. As the outcome was binary and time invariant (cognitive improvement vs no cognitive improvement; binominal distribution) a GEE model has more relaxed assumptions about the distribution of the data, and is therefore preferable (Adamis, 2009). Analysis indicated that the missing data were missing complete at random (MCAR) (Little’s MCAR test: Chisquare = 135.7, DF = 112, p = .07). 2.4. Ethics This study was approved by the local Research Ethics Committee of Guy’s and St Thomas’ Hospital, London, UK.

2. Materials and methods 3. Results The study was observational and longitudinal and was performed between July 2003 and April 2004 in the Elderly Care Unit of a University hospital serving an inner-city area. In this unit the admissions were from either; from home via the Accident & Emergency department or from other admission hospital wards if they needed specialist assessment Participants were consecutive patients aged 70 years or more assessed within 3 days of admission to an old age medical unit. Patients were excluded if they had

3.1. Demographic and baseline characteristics of the sample A total of 164 subjects participated in the study. Twenty two underwent a first assessment only and were thus excluded from further analyses. The mean age of the remaining sample (n = 142) was 84.8 (SD 6.4) and gender distribution comprized 47 males (33%) and 95 females. Nearly half (n = 64, 45%) had evidence of

D. Adamis et al. / Archives of Gerontology and Geriatrics 59 (2014) 593–598

595

Table 1 The means and medians of the continuous variables with asterisk (*) and in bold are those which significant different between the two groups. Improvement of MMSE by 20 percent and above No improvement

Age Frailty scale Barthel index DRS* APACHE II APS* IFN gamma pg/ml IL-1 alpha (pg/ml) IL-1 beta (pg/ml) IL-1 receptor alpha (pg/ml) IL-6 (pg/ml) TNF alpha (pg/ml) IGF-I* (ng/ml) LIF (pg/ml)

Improvement

Mean

Standard deviation

Median

Mean

Standard deviation

Median

84.94 1.73 13.58 6.55 9.73 3.37 5.18 .07 1.67 1433.69 17.23 5.30 80.52 4.29

7.12 1.01 4.58 4.91 2.84 2.74 8.75 .22 3.22 1255.69 12.40 6.69 44.03 7.42

84.84 2.00 14.00 5.00 9.00 3.00 .00 .00 .14 1074.50 16.07 4.51 74.53 .00

86.89 1.96 12.62 13.33 10.87 4.67 7.52 .07 1.02 1508.50 20.03 2.76 52.69 5.09

5.53 1.00 4.71 5.92 3.43 3.22 14.42 .16 2.98 968.53 19.26 4.05 18.32 9.38

87.84 2.00 13.00 13.00 11.00 4.00 .00 .00 .00 1386.20 15.82 .00 54.89 .00

dementia. Forty one (28.9%) patients had delirium at the first assessment (mean MMSE score 19.1  8.2). APOE testing was completed for 106 patients: 24 (23%) had at least one APOE epsilon 4 allele. Cytokines were analysed in the first 60 patients. 3.2. Correlates of cognitive improvement (bivariate analyses) 3.2.1. Cognitive improvement defined as a 20% improvement in MMSE score A score of 25 or less in MMSE is necessary to allow for a subsequent 20% improvement. The MMSE score was 25 or less in 105 patients/participants. From these, the 45 (43%) demonstrated an improvement in MMSE scores by 20% or more and the 60 (57%) who did not. Table 1 shows the demographic, biological and measured variables of the 105 participants at first assessment. Comparison of these two groups (improved versus did not improve) shows significant differences in relation to the APS, (Mann–Whitney U = 1035.5, p = .04), DRS score (Mann–Whitney U = 484, p < .001), IGF-I (Mann–Whitney U = 132.5, p = .04) and delirium status as defined by the CAM (x2 = 26.4, DF:1, p < .001). No significant differences were found in cytokine levels, age, APACHE II scores, frailty scale or Barthel index, nor were there significant differences for APOE epsilon 4 allele status, gender or previous history of dementia. In summary, subjects with more severe acute physical illness and more severe delirium at first assessment were more likely to improve cognitively compared to those with less severe physical illness and less severe delirium.

3.2.2. Cognitive improvement defined as an increase of at least 3 points in MMSE score This analysis was restricted to those with MMSE scores of 27 or less at first assessment (n = 123). Fifty two (42%) demonstrated an improvement by 3 points or more on the MMSE at a subsequent assessment, while the rest 71 (58%) did not. Table 2 shows the characteristics at first assessment of these two groups. Comparison of the two groups shows that those who improved were more likely to be delirious at first assessment according to CAM (x2 = 15.6, DF:1, p = .001), had more severe delirium according to the DRS (Mann–Whitney U = 948.5, p < .001), were significantly older (Mann–Whitney U = 1381, p = .02) were more likely to be female (x2 = 8.9, DF:1, p = .003) and with lower levels of circulating IGF-I (Mann–Whitney U = 156, p = .028). No significant differences were noted in respect of cytokine levels, frailty scale, Barthel index, the severity of physical illness (APACHE II and APS), APOE epsilon 4 allele, status or previous history of dementia. 3.3. Relationship between occurrence of delirium and changing cognitive status Using the 20% definition of improvement, 28 participants (53.8%) who experienced cognitive improvement had evidence of delirium, while for the definition of cognitive improvement requiring an increase of 3 or more points on the MMSE, 30 participants (65.2%) who experienced cognitive improvement had

Table 2 The means and medians of the continues variables with asterisk (*) and in bold are those which are significantly different between the two groups. Improvement on MMSE by 3 points and above No improvement

Age* Frailty scale Barthel index DRS* APACHE II APS IFN gamma (pg/ml) IL-1 alpha (pg/ml) IL-1 beta (pg/ml) IL-1 receptor alpha (pg/ml) IL-6 (pg/ml) TNF alpha (pg/ml) IGF-I* (ng/ml) LIF (pg/ml)

Improvement

Mean

Standard deviation

Median

Mean

Standard deviation

Median

84.33 2 14 6 10 4 5.34 .068 1.38 1202.15 15.12 5.41 80.44 6.83

6.71 1 4 5 3 3 8.94 .22 2.98 615.43 10.52 6.93 40.86 9.48

83.36 2 14 4 10 3 .00 .00 .06 1113.30 13.48 4.20 74.63 .00

86.87 2 13 11 10 4 6.42 .056 1.25 1631.18 20.45 2.98 55.58 3.21

5.82 1 5 7 3 3 12.99 .15 3.03 1358.74 18.25 3.94 23.24 6.46

88.24 2 14 12 10 4 .015 .00 .00 1288.47 16.41 .73 55.10 .00

D. Adamis et al. / Archives of Gerontology and Geriatrics 59 (2014) 593–598

596

Table 3 The estimates of the effects of the variables and their significance (in bold) during the hospitalization on cognitive improvement as it defined by either definition. The response category was improvement. 95% Wald confidence interval

Parameter

(Intercept) Male Female IL-1 alpha (pg/ml) IGF-I (ng/ml) Dementia [NO] Dementia [YES] No possession of e4 allele Possession of e4 allele IL-1 beta (pg/ml) No possession of e4 allele  IGF-I Possession of e4 allele  IGF-I Dementia [NO]  IGF-I Dementia [YES]  IGF-I

B

Std. error

.10 4.51 0a 6.34 .03 .61 0a 6.89 0a .18 .07 0a .08 0a

1.19 1.47 . 1.71 .02 1.73 . 2.05 . .07 .024 . .03 .

Hypothesis test

Lower

Upper

Wald Chi-square

DF

Sig.

2.44 7.39 . 9.68 .00 2.78 . 2.87 . .31 .12 . .13 .

2.24 1.62 . 2.99 .06 4.00 . 10.90 . .04 .03 . .03 .

.01 9.38 . 13.77 3.94 .13 . 11.32 . 6.01 9.08 . 8.79 .

1 1 . 1 1 1 . 1 . 1 1 . 1 .

.93 .002 . <.001 .047 .724 . .001 . .014 .003 . .003 .

Response category = improvement. The sign (+ or ) in the second column implies the direction of the relationship between the variable and cognitive improvement. E.g. lower IL-1 alpha was associated with greater chance of improvement at a significance level of p < .001 and negative estimate in categorical variables means that the other category is positive. E.g. negative estimate in males means that females are positively associated with the response category (improvement).

delirium. Using either definition of cognitive improvement from the 142 who underwent a second assessment, a total of 55 (38.7%) improved cognitively during the hospital stay, of these 30 (54.5%) had prevalent delirium. Similarly, ninety six participants (67.6%) never developed delirium during hospitalization and underwent continued follow-up. Of these, 23 (24%) improved cognitively by either the 20% or 3-point definitions. These patients had an overall mean DRS score of 4.5 (SD 2.8). 3.4. Longitudinal analysis of biological and phenomenological factors associated with cognitive improvement Using either of the two criteria for cognitive improvement, the data were analysed longitudinally using GEE in order to examine the relationship between variables and cognitive status (cognitive improvement or not) over time. In this model predictor variables included demographic characteristics, severity of illness, presence or absence of dementia, possession of APOE 4 allele, CAM status over time, DRS score, Cytokines and IGF-I levels. The number of observations was 133, involving 48 patients. The final most parsimonious model is presented in Table 3 which indicates how cognitive improvement was positively related to female gender, (lower) levels of circulating IL-1 alpha and IL-1 beta, (greater) levels of IGF-I and lack of APOE epsilon 4 allele. Also, the interaction of IGF- I levels with dementia status and possession of APOE epsilon 4 had a significant effect on cognitive improvement. More specifically, for patients with similar IGF-I levels, those with at least one APOE epsilon 4 allele were more likely to cognitively improve compared to those without. Similarly, for patients with similar IGF-I levels, those with dementia were more likely to improve compared to those without dementia. However, possession of APOE epsilon 4 allele remained a significantly strong independent risk factor for non-improvement. Neither delirium status nor delirium severity was significantly associated with cognitive improvement during hospitalization. 4. Discussion This study examines factors associated with improved cognitive function in hospitalized elderly medical patients. We did not make any assumptions as to the relevance of delirium or not, but rather measured how changing cognition related to a range of biological and other parameters, including expression of cytokines, level of physical morbidity, APOE e4 status, the presence or absence of

delirium, the presence or absence of dementia and demographics. We found that a variety of factors correlated with improving cognition and that, although substantial changes to cognitive performance occurred mostly in delirious patients, many subjects without delirium also experienced substantial clinically-relevant changes to cognition. This serves to emphasize that all acute and reversible cognitive disorder is not necessarily delirium. We used two separate criteria to define ‘cognitive improvement’ and found somewhat contrasting results whereby the presence of delirium, its severity and also IGF-I levels were significant factors using both criteria. The severity of illness only emerged as a significant factor associated with cognitive improvement when defined by a proportionate improvement of at least 20%. The merits of these two arbitrary criteria have not been externally validated – significant cognitive improvement has been equated with a two-point (Fields et al., 1986) or three point (Inouye et al., 2006) improvement in MMSE, while (Keeffe, Mulkerrin, Nayeem, Varughese, & Pillay, 2005) in a serial of MMSE tests suggest that 3 or more points improvement in MMSE can detect resolution of delirium (77% sensitivity, 75% specificity against CAM). Given that the MMSE is not a ratio scale (i.e. a score of 20 does not equate with twice the cognitive performance of a score of 10) with arbitrary anchor points (e.g. a score of 0 point does not mean that someone has no cognitive function at all), the criterion of improvement evidenced by changes in a specific number of points may artificially ‘‘penalize’’ those with low initial MMSE scores. The use of percentage changes in MMSE scores may be a better indicator because it accounts for the influence of baseline performance on subsequent change. In bivariate analysis cognitive improvement was predicted by both biological and phenomenological variables, such that subjects with more severe acute illness, lower levels of IGF-I and more severe delirium were more likely to improve by at least 20% in their MMSE scores. These findings are in keeping with the longstanding notion of delirium as a cognitive disorder that occurs in response to severe physical illness (Lipowski, 1990). However, a third of patients who improved cognitively did not have CAM-positive delirium, highlighting that cognitive difficulties in the physically ill are not entirely accounted for by delirium (Inouye et al., 2006). Similarly, longitudinal analyses did not identify delirium as a major factor in cognitive recovery. These findings emphasize the need to explore the relationship between non-delirious acute cognitive disorder in physically morbid patients and delirium. Key issues include comparisons of phenomenological profile, aetiological and

D. Adamis et al. / Archives of Gerontology and Geriatrics 59 (2014) 593–598

pathophysiological underpinnings, as well as therapeutic and prognostic implications. It is not known, for example, the extent to which non-delirious acute cognitive disorder is characterized by similar cognitive impairment with disproportionate inattention or if it overlaps with the concept of subsyndromal delirium. The relatively low mean DRS scores for this group suggest that the latter explanation does not account for a substantial number of cases. This work raises the issue of the relationship between biological changes that occur in delirious patients and the pathophysiology of delirium such that these may be an inherent part of the pathogenesis of delirious states or may be merely epiphenomenal. For example, post-operative delirium is characterized by reduced serum levels of amino acids (van der Mast & Fekkes, 2000) with tryptophan levels below 40 mg/ml suggested as a measure for delirium detection (Robinson, Raeburn, Angles, & Moss, 2008). However, this pattern has not been demonstrated in elderly medical patients (Flacker & Lipsitz, 2000) and it has been suggested therefore that altered amino acid levels may reflect the physiological response to surgery rather than a specific pathophysiological mechanism for delirium. Similarly, delirium symptoms overlap with features of sickness behavior that occurs in inflammatory responses (Holmes, Cunningham, Zotova, Culliford, & Perry, 2011). C-reactive protein is a marker of the acute phase of the inflammatory response that predicts delirium incidence in elderly medical admissions but lacks specificity for delirium and may be of greater use in monitoring progress (Macdonald, Adamis, Treloar, & Martin, 2007). Studies of cytokine levels and delirium incidence have been inconsistent, with some indicating elevated proinflammatory cytokines in elderly medical admissions (Rooij, van Munster, Korevaar, & Levi, 2007) with others suggesting reduced levels of anti-inflammatory and neuroprotective factors (Adamis & Meagher, 2011; Adamis et al., 2007; Wilson, Broadhurst, Diver, Jackson, & Mottram, 2005) while a study in surgical patients found elevated chemokine levels in the immediate post-operative period without a significant rise in cytokine levels (Rudolph et al., 2008). Cytokine and chemokines are associated with cognitive function and cholinergic activity such that their role as predisposing or precipitating factors or merely epiphenomena needs to be further explored. Previous studies have identified an association between cognitive impairment and low levels of IGF-I consistent with the recognized neuroprotective actions of IGF-I (Aleman et al., 1999; Arai et al., 2001). However, bivariate analysis indicated that lower levels of IGF-I were associated with subsequent increase of MMSE scores. As such, this may reflect a role for reduced IGF-I in the pathogenesis of more reversible causes of acute cognitive impairment. Similarly, longitudinal analysis indicated a positive association between increased levels of IGF-I and cognitive improvement after controlling for other variables. In addition, changes to IGF-I expression across time may vary depending on factors such as APOE and dementia status. The severity of physical illness as measured with APACHE II and APS did not have an independent effect on cognitive improvement throughout the admission although severity of illness was linked to greater likelihood of cognitive recovery in bivariate analyses. One explanation is that IGF-I levels are more sensitive indicators of severity of underlying illness than the APACHE II or APS measurements. In support, a reduction of IGF-I directly during acute inflammation or critical illness has been reported by Lackey and colleagues (Lackey, Gray, & Henricks, 2000). Similarly, in the longitudinal analysis, neither delirium nor its severity was identified as a factor in cognitive improvement. This may relate to the observation that some patients who appear recovered from delirium experience persistent cognitive deficits or so-called

597

long-term cognitive impairment (Adamis, Treloar, Martin, & Macdonald, 2006; Gross et al., 2012; MacLullich et al., 2009; Meagher, Adamis, Trzepacz, & Leonard, 2012). Longitudinal analysis found an association between higher levels of pro inflammatory cytokines IL-1 alpha and IL-1 beta and poor cognitive recovery as defined by either definition. Previous reports have also indicated that IL-1 alpha and IL-1 beta adversely impact upon cognitive function in both humans and animals (Cibelli et al., 2010; Holmes et al., 2009, 2003; Huang & Sheng, 2010; Yaffe et al., 2003). Both bivariate and longitudinal analysis suggested that females were more likely to experience cognitive improvement, perhaps reflecting the protective effect of oestrogens upon cognition, which is evident even after the menopause (Berent-Spillson et al., 2012; Daniel, 2013). 4.1. Limitations A limitation of the study is that patients were not followed up after discharge from hospital, so that more delayed changes to some measurements including cognition were not studied. Moreover, it is known that for many patients, including those with delirium, recovery can be a more gradual process, and as such the frequency of cognitive improvement is likely to have been underestimated in this population. The initial starting point from which improvement in MMSE was derived differs: Treloar and Macdonald used the lower score of MMSE of the two first assessments while (Inouye et al., 2006), and Fields et al. (1986) the first MMSE score by assuming this to be the lower). The diagnoses of dementia were based upon clinical history but would be more robust if this included formal assessment with tools such as the IQCODE that can identify dementia in patients with active delirium (Jorm, 2004). Finally, cytokines were measured in a relatively small subset of patients which limited the capacity to explore the relevance of factors such as separate cognitive items of MMSE, or of individual symptom domains to changes in biological parameters. 5. Conclusions This study examines the clinical correlates of cognitive improvement defined according to improved MMSE scores in medically ill older patients. Cognitive disturbance in physically ill elderly patients relates to a wide range of factors that should be considered when assessing likely causation and management. This work highlights the occurrence of non-delirious acute cognitive disorder, which warrants careful study in relation to phenomenological, etiological, and pathophysiological profile as well as therapeutic and prognostic implications as compared to delirium. Studies need to qualify the character and extent of the acute deterioration of cognition that equates with delirium since impairment of cognitive performance is common during periods of high morbidity or exposure to interventions that confer elevated delirium risk, including many who do not develop delirium. Conflict of interest None declared. References Adamis, D. (2009). Statistical methods for analysing longitudinal data in delirium studies. International Review of Psychiatry, 21(1), 74–85. http://dx.doi.org/ 10.1080/09540260802675346. pii:908710341. Adamis, D., & Macdonald, A. J. (2009). A review of the association of apolipoprotein E and delirium. In L. R. Penfield & R. T. Nelson (Eds.), Apoprotein research (pp. 41–77). New York: Nova Science Publishers. Adamis, D., & Meagher, D. (2011). Insulin-like growth factor I and the pathogenesis of delirium: A review of current evidence. Journal of Aging Research, 2011, 951403. http://dx.doi.org/10.4061/2011/951403

598

D. Adamis et al. / Archives of Gerontology and Geriatrics 59 (2014) 593–598

Adamis, D., Treloar, A., Gregson, N., Macdonald, A. J. D., & Martin, F. C. (2011). Delirium and the functional recovery of older medical inpatients after acute illness: The significance of biological factors. Archives of Gerontology and Geriatrics, 52, 276–280. Adamis, D., Treloar, A., Martin, F. C., Gregson, N., Hamilton, G., & Macdonald, A. J. (2007). APOE and cytokines as biological markers for recovery of prevalent delirium in elderly medical inpatients. International Journal of Geriatric Psychiatry, 22(7), 688– 694. http://dx.doi.org/10.1002/gps.1732 Adamis, D., Treloar, A., Martin, F. C., & Macdonald, A. J. (2006). Recovery and outcome of delirium in elderly medical inpatients. Archives of Gerontology and Geriatrics, 43(2), 289–298. http://dx.doi.org/10.1016/j.archger.2005.11.005. pii:S01674943(05)00156-1. Aleman, A., Verhaar, H. J., De Haan, E. H., De Vries, W. R., Samson, M. M., Drent, M. L., et al. (1999). Insulin-like growth factor-I and cognitive function in healthy older men. Journal of Clinical Endocrinology and Metabolism, 84(2), 471–475. Arai, Y., Hirose, N., Yamamura, K., Shimizu, K., Takayama, M., Ebihara, Y., et al. (2001). Serum insulin-like growth factor-1 in centenarians: Implications of IGF-1 as a rapid turnover protein. Journals of Gerontology Series A: Biological Sciences and Medical Sciences, 56(2), M79–M82. Avidan, M. S., & Evers, A. S. (2011). Review of clinical evidence for persistent cognitive decline or incident dementia attributable to surgery or general anesthesia. Journal of Alzheimer’s Disease, 24, 201–216. Berent-Spillson, A., Persad, C. C., Love, T., Sowers, M., Randolph, J. F., Zubieta, J. K., et al. (2012). Hormonal environment affects cognition independent of age during the menopause transition. Journal of Clinical Endocrinology and Metabolism, 97(9), E1686–E1694. http://dx.doi.org/10.1210/jc.2012-1365. pii:jc.2012-1365. Cibelli, M., Fidalgo, A. R., Terrando, N., Ma, D., Monaco, C., Feldmann, M., et al. (2010). Role of interleukin-1beta in postoperative cognitive dysfunction. Annals of Neurology, 68(3), 360–368. http://dx.doi.org/10.1002/ana.22082 Cole, M. G., You, Y., McCusker, J., Ciampi, A., & Belzile, E. (2008). The 6 and 12 month outcomes of older medical inpatients who recover from delirium. International Journal of Geriatric Psychiatry, 23(3), 301–307. http://dx.doi.org/10.1002/gps.1878 Daniel, J. M. (2013). Estrogens, estrogen receptors, and female cognitive aging: The impact of timing. Hormones and Behavior, 63(2), 231–237. http://dx.doi.org/ 10.1016/j.yhbeh.2012.05.003. pii:S0018-506X(12)00144-4. de Rooij, S. E., van Munster, B. C., Korevaar, J. C., & Levi, M. (2007). Cytokines and acute phase response in delirium. Journal of Psychosomatic Research, 62(5), 521–525. http://dx.doi.org/10.1016/j.jpsychores.2006.11.013. pii:S0022-3999(06)00526-5. Fields, S. D., MacKenzie, C. R., Charlson, M. E., & Perry, S. W. (1986). Reversibility of cognitive impairment in medical inpatients. Archives of Internal Medicine, 146(8), 1593–1596. Flacker, J. M., & Lipsitz, L. A. (2000). Large neutral amino acid changes and delirium in febrile elderly medical patients. Journals of Gerontology Series A: Biological Sciences and Medical Sciences, 55(5), B249–B252. discussion B253-244. Gross, A. L., Jones, R. N., Habtemariam, D. A., Fong, T. G., Tommet, D., Quach, L., et al. (2012). Delirium and long-term cognitive trajectory among persons with dementia. Archives of Internal Medicine, 172(17), 1324–1331. http://dx.doi.org/10.1001/ archinternmed.2012.3203. pii:1351352. Holmes, C., Cunningham, C., Zotova, E., Culliford, D., & Perry, V. H. (2011). Proinflammatory cytokines, sickness behavior, and Alzheimer disease. Neurology, 77(3), 212–218. Holmes, C., Cunningham, C., Zotova, E., Woolford, J., Dean, C., Kerr, S., et al. (2009). Systemic inflammation and disease progression in Alzheimer disease. Neurology, 73(10), 768–774. http://dx.doi.org/10.1212/WNL.0b013e3181b6bb95. pii:73/10/ 768. Holmes, C., El-Okl, M., Williams, A. L., Cunningham, C., Wilcockson, D., & Perry, V. H. (2003). Systemic infection, interleukin 1beta, and cognitive decline in Alzheimer’s disease. Journal of Neurology, Neurosurgery & Psychiatry, 74(6), 788–789. Huang, Z. B., & Sheng, G. Q. (2010). Interleukin-1beta with learning and memory. Neuroscience Bulletin, 26(6), 455–468. http://dx.doi.org/10.1007/s12264-010-6023-5 Inouye, S. K., Zhang, Y., Han, L., Leo-Summers, L., Jones, R., & Marcantonio, E. (2006). Recoverable cognitive dysfunction at hospital admission in older persons during acute illness. Journal of General Internal Medicine, 21(12), 1276–1281. http:// dx.doi.org/10.1111/j.1525-1497.2006.00613.x. pii:JGI613. Jones, C., Griffiths, R. D., Slater, T., Benjamin, K. S., & Wilson, S. (2006). Significant cognitive dysfunction in non-delirious patients identified during and persisting

following critical illness. Intensive Care Medicine, 32(6), 923–926. http://dx.doi.org/ 10.1007/s00134-006-0112-y Jorm, A. F. (2004). The Informant Questionnaire on cognitive decline in the elderly (IQCODE): A review. International Psychogeriatrics, 16(3), 275–293. Lackey, B. R., Gray, S. L., & Henricks, D. M. (2000). Does the insulin-like growth factor system interact with prostaglandins and proinflammatory cytokines during neurodegeneration? Proceedings of the Society for Experimental Biology and Medicine, 224(1), 20–27. Lipowski, Z. J. (1990). Delirium: Acute confusional states. New York: Oxford University Press. Lowery, D. P., Wesnes, K., Brewster, N., & Ballard, C. (2008). Quantifying the association between computerised measures of attention and confusion assessment method defined delirium: A prospective study of older orthopaedic surgical patients, free of dementia. International Journal of Geriatric Psychiatry, 23(12), 1253–1260. http:// dx.doi.org/10.1002/gps.2059 Macdonald, A., Adamis, D., Treloar, A., & Martin, F. (2007). C-reactive protein levels predict the incidence of delirium and recovery from it. Age and Ageing, 36(2), 222– 225. http://dx.doi.org/10.1093/ageing/afl121. pii:afl121. MacLullich, A. M., Beaglehole, A., Hall, R. J., & Meagher, D. J. (2009). Delirium and longterm cognitive impairment. International Review of Psychiatry, 21(1), 30–42. http:// dx.doi.org/10.1080/09540260802675031. pii:908707924. Marx, F., Blasko, I., & Grubeck-Loebenstein, B. (1999). Mechanisms of immune regulation in Alzheimer’s disease: A viewpoint. Archivum Immunologiae et Therapiae Experimentalis, 47, 205–209. Meagher, D., Adamis, D., Trzepacz, P., & Leonard, M. (2012). Features of subsyndromal and persistent delirium. British Journal of Psychiatry, 200, 37–44. http://dx.doi.org/ 10.1192/bjp.bp.111.095273 pii:bjp.bp.111.095273. O’Keeffe, S. T., Mulkerrin, E. C., Nayeem, K., Varughese, M., & Pillay, I. (2005). Use of serial Mini-Mental State Examinations to diagnose and monitor delirium in elderly hospital patients. Journal of the American Geriatrics Society, 53(5), 867–870. http:// dx.doi.org/10.1111/j.1532-5415.2005.53266.x. pii:JGS53266. Robinson, T. N., Raeburn, C. D., Angles, E. M., & Moss, M. (2008). Low tryptophan levels are associated with postoperative delirium in the elderly. American Journal of Surgery, 196(5), 670–674. http://dx.doi.org/10.1016/j.amjsurg.2008.07.007. pii:S0002-9610(08)00560-6. Rudolph, J. L., Ramlawi, B., Kuchel, G. A., McElhaney, J. E., Xie, D., Sellke, F. W., et al. (2008). Chemokines are associated with delirium after cardiac surgery. Journals of Gerontology Series A: Biological Sciences and Medical Sciences, 63(2), 184–189. pii:63/ 2/184. Simone, M. J., & Tan, Z. S. (2011). The role of inflammation in the pathogenesis of delirium and dementia in older adults: A review. CNS Neuroscience & Therapeutics, 17(5), 506–513. http://dx.doi.org/10.1111/j.1755-5949.2010.00173.x. pii:CNS173. Torgersen, J., Hole, J. F., Kvale, R., Wentzel-Larsen, T., & Flaatten, H. (2011). Cognitive impairments after critical illness. Acta Anaesthesiologica Scandinavica, 55(9), 1044– 1051. http://dx.doi.org/10.1111/j.1399-6576.2011.02500.x Treloar, A. J., & Macdonald, A. J. (1997a). Outcome of delirium. Part 1. Outcome of delirium diagnosed by DSM-III-R, ICD-10 and CAMDEX and derivation of the Reversible Cognitive Dysfunction Scale among acute geriatric inpatients. International Journal of Geriatric Psychiatry, 12(6), 609–613. 10.1002/(SICI)10991166(199706)12:6<609::AID-GPS553>3.0.CO;2-L [pii]. Treloar, A. J., & Macdonald, A. J. (1997b). Outcome of delirium. Part 2. Clinical features of reversible cognitive dysfunction – Are they the same as accepted definitions of delirium? International Journal of Geriatric Psychiatry, 12(6), 614–618. 10.1002/ (SICI)1099-1166(199706)12:6<614::AID-GPS1553>3.0.CO;2-X [pii]. van der Mast, R. C., & Fekkes, D. (2000). Serotonin and amino acids: Partners in delirium pathophysiology? Seminars in Clinical Neuropsychiatry, 5(2), 125–131. http:// dx.doi.org/10.153/SCNP00500125. pii:00500125. Wilson, K., Broadhurst, C., Diver, M., Jackson, M., & Mottram, P. (2005). Plasma insulin growth factor-1 and incident delirium in older people. International Journal of Geriatric Psychiatry, 20(2), 154–159. http://dx.doi.org/10.1002/gps.1265 Wilson, M. M., & Morley, J. E. (2003). Impaired cognitive function and mental performance in mild dehydration. European Journal of Clinical Nutrition, 57(Suppl. 2), S24– S29. http://dx.doi.org/10.1038/sj.ejcn.1601898. pii:1601898. Yaffe, K., Lindquist, K., Penninx, B. W., Simonsick, E. M., Pahor, M., Kritchevsky, S., et al. (2003). Inflammatory markers and cognition in well-functioning African-American and white elders. Neurology, 61(1), 76–80.