Persistent delirium in older hospital patients: a systematic review of frequency and prognosis

Transcription

1 Age and Ageing 2009; 38: C The Author Published by Oxford University Press on behalf of the British Geriatrics Society. doi: /ageing/afn253 All rights reserved. For Permissions, please journals.permissions@oxfordjournals.org Published electronically 18 November 2008 SYSTEMATIC REVIEW Persistent in older hospital patients: a systematic review of frequency and prognosis MARTIN G. COLE 1,2,ANTONIO CIAMPI 3,4,ERIC BELZILE 3,LIHONG ZHONG 3 1 Department of Psychiatry, St. Mary s Hospital Center, Montreal, Quebec, Canada 2 Department of Psychiatry, McGill University, Montreal, Quebec, Canada 3 Department of Clinical Epidemiology and Community Studies, St. Mary s Hospital Center, Montreal, Quebec, Canada 4 Department of Epidemiology and Biostatistics, McGill University, Montreal, Quebec, Canada Address correspondence to: Martin G. Cole. Tel: (+1) , Ext. 5060; Fax: (+1) martin.cole@ssss.gouv.qc.ca Abstract Background: one explanation for the poor prognosis of among older hospital patients may be that many of these patients do not recover from. We sought to determine the frequency and prognosis of persistent (PerD) in older hospital patients by systematically reviewing original research on this topic. Methods: MEDLINE, EMBASE, PsycINFO and the Cochrane Database of Systematic Reviews were searched for potentially relevant articles. The bibliographies of relevant articles were searched for additional references. Eighteen reports (involving 1,322 patients with ) met the following seven inclusion criteria: original research published in English or French, prospective study design, study population of at least 20 hospital patients, patients aged 50 years or more, follow-up of at least 1 week, acceptable definition of at enrolment and included at least one assessment for PerD at discharge or later. The methods of each study were assessed according to the six criteria for prognostic studies described by the Evidence-Based Medicine Working Group. Information about the sample origin and size, age, proportion with dementia, criteria for, timing of follow-up assessments, criteria for PerD, proportion with PerD and prognosis of PerD was systematically abstracted from each report, tabulated and combined using standard meta-analysis techniques. Results: the combined proportions with PerD at discharge, 1, 3 and 6 months were 44.7% (95% CI 26.8%, 63.7%), 32.8% (95% CI 18.4%, 47.2%), 25.6% (95% CI 7.9%, 43.4%) and 21% (95% CI 1.4%, 40.6%), respectively. The outcomes (mortality, nursing home placement, function, cognition) of patients with PerD were consistently worse than the outcomes of patients who had recovered from. Conclusion: PerD in older hospital patients is frequent, appears to be associated with adverse outcomes and may account for the poor prognosis of in this population. These findings have potentially important implications for clinical practice and research. Keywords: persistent, aged, frequency, prognosis, elderly Introduction Delirium is a cognitive disorder defined by acute onset, fluctuating course and disturbances of consciousness, attention, orientation, memory, thought, perception and behaviour [1]. It occurs in hyperactive, hypoactive or mixed forms in up to 50% of older hospital patients [2,3], many with pre-existing dementia [4]. Traditionally, the course of has been described as transient [5], in which recovery is likely to be complete if the underlying etiological factor is promptly corrected or self-limited [1]. Among older hospital patients, however, the prognosis is poor [3,6]. In this population, is associated with significant increases in cognitive impairment and functional disability [7 9], length of hospital stay [10,11], rates of institutionalisation [7,9] and rates of death [12 14], independent of many socio-demographic (e.g. age, gender, marital status, living arrangements) and clinical (e.g. presence of dementia, co-morbidity, illness severity) variables. 19

2 M. G. Cole et al. One explanation for the poor prognosis of among older hospital patients may be that many of these patients do not recover from [15,16] and that the persistence of, rather than the occurrence of an episode of per se, accounts for many of the adverse outcomes. Because this issue has potentially important implications for clinical practice and research, the primary objectives of this study were to determine the frequency and prognosis of persistent (PerD) in older hospital patients by systematically reviewing original research on this topic. For the purpose of this review, persistent was defined as a cognitive disorder that met accepted diagnostic criteria for at admission (or shortly after admission) and continued to meet criteria for at the time of discharge or beyond. The review process, modified from the one described by Oxman et al. [17], involved systematic selection of articles, assessment of validity, abstraction of data and qualitative and quantitative synthesis of results. Methods Selection of articles The selection process involved four steps. First, four computer databases, MEDLINE, EMBASE, PsycINFO and the Cochrane Database of Systematic Reviews, were searched for potentially relevant articles published from January 1966 to September 2007, January 1996 to September 2007, January 1985 to September 2007 and to the third quarter of 2007, respectively. The key words (title, abstract, text, key words, subject headings, identifiers) used for all databases were, prognosis, outcome and aged. Second, relevant articles (based on the title and abstract) were retrieved for more detailed evaluation. Third, the bibliographies of relevant articles were searched for additional references. Finally, all retrieved articles were screened to meet the following seven inclusion criteria: (i) original research published in English or French; (ii) study population of at least 20 (current or recent) hospital patients; (iii) patients aged 50 years or more; (iv) prospective study; (v) follow-up of at least 1 week; (vi) acceptable definition of at enrolment; (vii) included at least one assessment for PerD at discharge or later. Assessment of validity To determine validity, the methods of each study were assessed according to the six criteria for prognostic studies described by the Evidence-Based Medicine Working Group [18]: (i) representative and well-defined sample of patients at a similar point in the course of the disease; (ii) follow-up that was sufficiently long (i.e. 3 months); (iii) follow-up that was complete (i.e. 80% of the inception cohort); (iv) use of objective outcome criteria; (v) unbiased outcome assessment (i.e. blind to inception diagnosis of ); (vi) adjustment for potentially confounding prognostic factors (e.g. dementia, severity of medical illness). Each study was scored with respect to meeting (+) or not meeting ( ) each of the above criteria. Abstraction of data Information about the sample origin and size, age, proportion with dementia, criteria for, timing of follow-up assessments, criteria for PerD, proportion with PerD and prognosis of PerD was systematically abstracted from each report and tabulated. Data synthesis (a) Qualitative: a qualitative meta-analysis was conducted by summarising, comparing and contrasting abstracted data. (b) Quantitative: standard meta-analysis techniques were applied to different groups of studies. Meta-analysis was used to calculate the combined estimate of prevalence of PerD at different time points. According to usual practice we used a fixed effect model first, followed by a test of homogeneity. If the test of homogeneity was rejected, the random effect model was used [19]. All analyses were performed using STATA software Results Selection of articles The search strategy yielded 250 potentially relevant studies; 46 were retrieved for more detailed evaluation. Eighteen studies met the inclusion criteria (Table 1) [10,15,20 35]. The other 28 studies were excluded for the following reasons: 25 did not include at least one assessment for PerD at discharge or later and 3 did not meet two or more of the inclusion criteria. Assessment of validity The results of the validity assessment are presented in Table 2. Three studies met all of the criteria. Many studies did not have a follow-up that was sufficiently long, conduct unbiased outcome assessments or adjust for potentially confounding prognostic factors. Data synthesis (a) Qualitative: the 18 studies enrolled 1,322 patients with (Table 1). One study [35] included two cohorts, development and validation cohorts, but the results of the validation cohort were reported in an earlier publication [33]; consequently, we used only data from the development cohort. Patients were enrolled from hospital medical units in 13 studies, hospital medical and surgical units in 2 studies, hospital surgical units in 2 studies and post-acute care facilities in 1 study. Four studies enrolled only patients with prevalent ( present on admission), 4 studies enrolled only patients with incident ( developing after admission) and 10 studies enrolled patients with prevalent and incident. Subjects mean ages were years, median 82 years. The mean length of hospital stay ranged from 7 to 43 days, median 18 days. The length of reported follow-up ranged from discharge to 12 months, median 20

3 Persistent in older hospital patients Table 1. Summary of studies examining the frequency of persistent Frequency of persistent Author/year Patient population Mean (median) length of hospital stay (days) Prevalent (P) or incident (I) Number of patients with Mean age Patients with Criteria for dementia (%) Criteria for persistent Discharge 1 month 3 months 6 months Hodkinson/1973 M P, I Acute onset of confusion Increase in Mental Test score <5points Rockwood/1989 M 17 P, I DSM3 Clinical impression 0 Lev koff et al./1992 M-S 19 P, I DSM3 Presence of any DSM symptom of Williams et al./1992 S 16 I DSM3R Clinical impression 0 Jitapunkul et al./1992 M (20) P, I DSM3R Increase in abbreviated Mental Test score <2 points Rockwood/1993 M 32 P, I (est) 1 DSM3, DSM3R Presence of any DSM symptom of Gaudet et al./1993 M 43 P,l DSM3R Met DSM3R criteria for Rudberg et al./1997 M-S 15 P, I (est) 5 DSM3R Met DSM3R criteria for O Keeffe et al./1997 M 21 P, I ,7 DSM3 Met DSM3 criteria for Treloar et al./1997 M 21 P, I DSM3R, ICD-10, CAMDEX Increase in MMSE score <5 points Marcantonio et al./2006 S 7 I CAM Met CAM criteria for Kelly et al./2001 M 8 P CAM Met CAM criteria for McCusker et al./2003 M 18 P, I DSM3R Met Dl (DSM3R) criteria for Kiely et al./2004 PAC P CAM Met CAM criteria for Lundstrom et al./2005 M 16 P 62 (control group of CT) OBS scale Met OBS scale and concensus criteria for McAvay et al./2006 M 15 I CAM Met CAM criteria for Pitkala et al./2006 M 26 P 87 (control group of CT) DSM4 Decrease in MDAS score <4 points Inouye et al./2007 M 10 I ,7 CAM Met CAM criteria for (at 7 days) months (at 7 days) (at 8 days) M, medical units; S, surgical units; PAC, post-acute care facility; CT, clinical trial; est, estimated; DSM, Diagnostic and Statistical Manual of the American Psychiatric Association; ICD, International Classification of Diseases; CAMDEX, Cambridge Examination for Mental Disorders of the Elderly; OBS, organic brain syndrome; CAM, Confusion Assessment Method; MMSE, Mini-Mental State Exam; DI, Index; MDAS, Memorial Delirium Assessment scale. Criteria for dementia. 1 Not specified; 2 clinical impression; 3 DSM; 4 chart review; 5 MMSE < 21 24; 6 IQCODE > 3.5; 7 Blessed Dementia Scale = 4. 21

4 M. G. Cole et al. Table 2. Validity of prognostic studies with respect to meeting (+) or not meeting criteria derived from the Evidence-Based Medicine Working Group Formation of Follow-up Follow-up Use of objective Unbiased outcome Adjustment for Author/year inception cohort sufficiently long complete outcome criteria assessment prognostic factors... Hodkinson/ ± Rockwood/1989 ± + + ± Levkoff et al./1992, Williams et al./ ± Jitapunkul/ ± Rockwood/ Gaudet et al./ ? Rudberg et al./ ± O Keeffe et al./ Treloar et al./ Marcantonio et al./ Kelly et al./ ? McCusker et al./ Kiely et al./ Lundstrom et al./2005 +? + + McAvay et al./ Pitkala et al./ Inouye et al./ weeks. Criteria for dementia varied from one study to the next (see Table 1); the proportion of patients with dementia ranged from 0% to 75%, median 44%. Only three studies measured depressive symptoms at baseline [23,34,35]. Ten studies used DSM criteria to diagnose at enrolment, five used the Confusion Assessment Method (CAM) algorithm, one used the Organic Brain Syndrome (OBS) scale, one used the criterion of acute onset of confusion and one study used three different sets of criteria. Nine studies used validated diagnostic criteria for to determine the presence of PD: five used the CAM algorithm, three used DSM criteria and one used the OBS scale. The other nine studies used non-validated criteria: four used change scores on cognitive or rating scales, two used a clinical impression, two used the presence of any DSM symptom of and one used a DSM3R algorithm based on a rating scale. The proportions with PerD are presented by study in Table 1. s determined at 7 8 days after enrolment were included under proportions at discharge. s among studies using validated criteria for ranged from 1 to 60% (median, 39%) at discharge, 31 to 51% (median, 32 39%) at 1 month and 13 to 48% at 3 months (median, 18%). s among studies using nonvalidated criteria ranged from 0 to 78% (median, 32 54%) at discharge, 6 to 39% at 1 month, 18 to 48% at 3 months, 11 to 31% at 6 months; one study reported a proportion of 41% at 12 months. The proportions appeared to decline from discharge to 3 months but were still high at 6 and 12 months. Differences in the proportions from one study to the next could not be explained by differences in the patient population, mean age, proportion of patients with dementia, criteria for at enrolment, enrolment of incident or prevalent cases of or the use of validated criteria for the assessment of PerD; however, proportions with PerD at discharge were higher when the criterion was the presence of any DSM symptom of (i.e %) and lowest when the criterion was a clinical impression of no (i.e. 0%). Even though all included studies were studies of the prognosis of, only three studies (involving 288 patients with ) determined the outcomes of patients with PerD (Table 3). The first study [29] reported that PerD at 1 month (compared to patients whose had resolved by 1 month) was associated with increased mortality and nursing home placement and decreased function at 1 month, although the differences between groups failed to reach levels of clinical or statistical significance. The second study [15] reportedthatperdat24h(comparedtorecovered by 24 h) was independently associated with a clinically significant increase in mortality at 12 months and a clinically and statistically significant decline in function at 12 months; similarly, PerD at discharge (compared to recovered by 24 h) was independently associated with a clinically significant increase in mortality at 12 months and clinically and statistically significant declines in both cognition and function at 12 months. The third study [33] reported that PerD at discharge (compared to recovered by discharge) was independently associated with a clinically, but not statistically, significant increase in mortality and nursing home placement at 12 months. (b) Quantitative: Estimates of the proportions and combined proportions of PerD are presented in Figure 1. Combined proportions were 17.4% (95% CI 15.3%, 19.4%) at discharge, 32.8% (95% CI 18.4%, 47.2%) at 1 month, 25.6% (95% CI 7.9%, 43.4%) at 3 months and 21% (95% CI 1.4%, 40.6%) at 6 months; one study reported a proportion of 41% (95% CI 33.8%, 48.2%) at 12 months. Notably, when two studies with 0 values (and very narrow confidence intervals) at discharge were excluded from the analysis, there was a 22

5 Table 3. Summary of studies of the prognosis of patients with persistent Author/year N PerD determined at with PerD (%) Outcomes determined at Persistent in older hospital patients Outcomes [OR (95% CI)] Mortality NH placement ADL decline Cognitive decline Ambulation decline... Marcantonio et al./ month 32 1 month 1.5 (0.45, 5.0) a,e 3.05 (0.74, 5.73 (1.51, 12.64) a 21.78) a McCusker et al./ h after months 1.62 (0.79, 10.7 ( 19.8, 0.87 ( 1.1, enrolment 3.34) b 1.6) c 2.8) d Discharge months 1.63 (0.77, 21.9 ( 31.1, 5.3 ( 7.5, 3.44) b 12.7) c 3.1) d McAvay et al./ Discharge months 2.38 (0.64, 8.84) a,e NH, nursing home; ADL, activities of daily living; PerD, persistent. a Unadjusted odds ratio (95% confidence interval), b adjusted odds ratio, c adjusted mean difference on Barthel Index, d adjusted mean difference on Mini-Mental State Exam, e combined OR for death/nursing Home placement = 1.85 (0.76, 4.5). Rockwood_89 Levkoff_92 Williams_92 Rockwood_93 Gaudet_93 O'Keefe_97 Rudberg_97 Kelly_01 McCusker_03 Pitkala_04 Lundstrom_05 McAvay_06 Marcantonio_06 Inouye_07 Combined Combined* Discharge Hodkinson_73 Jitapunkul_92 Rudberg_97 Kelly_01 Kiely_04 Marcantonio_06 Combined 1 month Levkoff_92 3 months Levkoff_92 6 months Treloar_97 McCusker_03 Kelly_01 Combined Combined Figure 1. Individual and combined proportions (and 95% confidence intervals) for persistent at discharge and 1, 3, 6 months after enrolment. Combined estimate after excluding two studies with 0 values. 23

6 M. G. Cole et al. striking difference in the combined proportion at discharge, 44.7% (95% CI 25.8%, 63.7%). This latter proportion was considered to be a better summary of the discharge data presented in the studies. The estimate of one combined outcome OR, death or nursing home placement, was 1.85 (95% CI 0.76, 4.50). Estimates of the other combined ORs could not be calculated because of the differences in outcome measures from one study to the next. Discussion This study proposed to determine the frequency and prognosis of PerD in older hospital patients by systematically reviewing original research on this topic. The combined proportions with PerD at discharge, 1, 3 and 6 months were 44.7, 32.8, 25.6 and 21%, respectively; one study reported a proportion of 41% at 12 months. In three studies, the longer term outcomes (mortality, nursing home placement, cognition, function) of patients with PerD were consistently worse than the outcomes of patients who had recovered from, although the clinically significant differences between groups in many studies often failed to reach levels of statistical significance, perhaps because of the small numbers of subjects. It appears that many older hospital patients do not recover from and that the persistence of is associated with adverse outcomes. Thus, the inclusion of many patients with PerD in previous studies of prognosis [3,6 14] may account, at least in part, for many of the adverse outcomes reported to be associated with in this population. This possibility is supported by the findings of two recent studies. The first study [36] reported that the 6 and 12 month outcomes of patients with persistent subsyndromal (SSD) were poorer than the outcomes of patients who recovered from SSD. The second study [37] reported that most of the 6- and 12-month outcomes of patients who recovered from by 8 weeks and survived were similar to the outcomes of patients who did not have an index episode. The findings of this systematic review have five potentially important implications for clinical practice and research. First, the course and outcome of in older hospital patients must be re-examined in original studies that determine the outcomes of patients with recovered or PerD separately. Second, it appears that half of the patients with PerD at discharge had recovered by 3 months. It is unknown whether this delayed recovery has adverse prognostic implications although one study included in the review [15] reported that faster recovery from was associated with better outcomes. Third, it is unknown whether most of the patients with PerD at 3 months can ever recover, even with intervention. There should be efforts to identify patients with PerD at 3 months in order to explore the usefulness of different types of interventions. The presence of persistent symptoms of suggests that these interventions might involve protocols to detect and manage putative causes of PerD such as unresolved medical illness or unrecognised drug toxicity. To date, the results of intervention trials that have targeted all patients with on admission to hospital or shortly after admission have been modest [32,34,38] or negative [39], possibly because these trials included many patients with transient who recovered with usual care, irrespective of any intervention effect. Alternatively, these trials may have included many patients with PerD. Trials that target patients with PerD at discharge or 3 months and use interventions specific to PerD may demonstrate a greater impact on outcomes. Fourth, the persistence of and associated cognitive impairment may interfere with the patient s selfmanagement of chronic medical conditions (e.g. poor compliance with medication). This impaired self-management may, in turn, contribute to adverse outcomes. Finally, possible relationships between PerD and functional decline, mild cognitive impairment and new-onset dementia should be explored. Dementia is the strongest risk factor for among older patients [40]; appears to increase the risk of dementia [13]. Four studies included in the review reported that the presence of dementia was the strongest risk factor for PerD [15,31,33,35] Many patients with dementia appear to have some symptoms of such as inattention and fluctuation [41]. Both and dementia are characterised by reduced metabolic rates and impaired cholinergic function [42]. Many conditions that cause can also cause dementia if they are prolonged and severe (e.g. hypoxia, hypoglycaemia), presumably related to excitotoxic damage and death of neurons [43]. Interestingly, prior to the twentieth century, persistence of symptoms and possible progression to dementia were considered common features of [44]. The studies of PerD included in this systematic review have three potential limitations. First, many studies used nonvalidated criteria to determine proportions with PerD; however, the proportions with PerD were similar in studies using validated or non-validated criteria (Table 1). Second, most of the studies enrolled patients with prevalent ( present on admission). Many patients may have had PerD long before admission and represent prevalent PerD rather than incident PerD; however, the frequency and prognosis of PerD did not seem to be related to enrolment of patients with prevalent as opposed to incident (Table 1). Third, the high proportions of after discharge (i.e. at 1, 3, 6 and 12 months) suggest that PerD persists well beyond discharge; however, because the reported proportions after discharge were group proportions derived from assessments repeated after intervals of weeks or months, it was not possible to determine if the high proportions of after discharge reflect PerD in the same patients or episodes of recurrent in different patients. Consequently, it is not possible to determine whether the adverse outcomes associated 24

7 Persistent in older hospital patients with PerD at discharge are related to the presence of an ongoing process or to recurrent episodes of. This systematic review has three potential limitations. First, the literature search was limited to articles published in English and French because resources to translate articles written in other languages were not available. Second, publication bias was not assessed, although it is unlikely that this bias influences publication of studies of frequency or prognosis. Third, relatively few studies determined the prognosis of PerD and the examination of prognosis was complicated by the small numbers of subjects and differences in the criteria for PerD, length of follow-up and outcome measures from one study to the next. Conclusion PerD in older hospital patients is frequent, appears to be associated with adverse outcomes and may account for the poor prognosis of in this population. Despite the limitations of the studies of PerD and this systematic review, these findings suggest that PerD warrants increased attention from clinicians and researchers. Recommendations The following recommendations should be considered in the design of new studies to characterise the phenomenology, predictors, course and outcome of PerD. Selection criteria should include explicit diagnostic criteria for and PerD; criteria for PerD should include the time of assessment for PerD; patients with imminently terminal illnesses should be excluded. s and outcomes of PerD in cohorts of patients with prevalent and incident should be determined separately. Patients should be assessed at least every few days throughout the follow-up period to determine the evolution of PerD and whether these disorders are recurrent rather than persistent or a combination of both. Phenomenology (including good measures of cognition and mood) should be assessed using one or more widely accepted instruments. It is especially important to develop and use better clinical measures of level of consciousness, attention and psychomotor activity. Important characteristics of subjects, potential risk factors and possible confounding factors (e.g. dementia, type and severity of physical illnesses, functional status, pre-admission institutional status and medication use) should be assessed using reliable and valid measures. Subjects could be categorised by the presence or absence of dementia, and the prognosis of each category could be reported separately. The minimum follow-up period should be 6 months, and the follow-up period should begin at enrolment. Outcome categories should include cognition, activities of daily living and rates of new-onset dementia, institutional care and mortality. The longer term outcomes of patients with PerD at different time points (e.g. discharge and 1, 3 and 6 months after discharge) should be compared to both the outcomes of patients who recover from at different time points and the outcomes of patients who did not have an index episode of. The assessment of outcomes should be independent of the assessments for PerD and the prognostic factors for PerD. Outcomes should be adjusted for potential confounders (e.g. dementia, severity of medical illness). Comparison groups should include patients with recovered prevalent and incident and no. Procedures to detect and manage should be recorded in detail and related to frequency of PerD; similarly, procedures to detect and manage PerD should be recorded in detail and related to outcomes. All phases of the study (e.g. selection of cases, diagnosis and outcome measures) should have demonstrated reliability. Key points Persistent is frequent in older hospital patients. Patients with persistent appear to have worse outcomes than patients who recover from. Inclusion of patients with persistent in previous studies of the prognosis of may account, at least in part, for the adverse outcomes reported to be associated with in older hospital patients. Use of protocols to detect and treat that persists at discharge or later may improve the longer term outcomes of in older hospital patients. Conflicts of interest None declared. References 1. American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders, 4th edition. Washington, DC: American Psychiatric Association, Levkoff S, Cleary P, Liptzin B, Evans DA. Epidemiology of : an overview of research issues and findings. Int Psychogeriatr 1991; 3: Siddiqi N, House AO, Holmes JD. Occurrence and outcome of in medical in-patients: a systematic literature review. Age Ageing 2006; 35(4): Fick DM, Agostini JV, Inouye SK. Delirium superimposed on dementia: a systematic review. J Am Geriatr Soc 2002; 50: American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders. 3rd edition. Washington, DC: American Psychiatric Association,

8 M. G. Cole et al. 6. Cole M, Primeau F. Prognosis of in elderly hospital patients. Can Med Assoc J 1993; 149: McCusker J, Cole MG, Dendukuri N, Belzile E, Primeau F. Delirium in older medical inpatients and subsequent cognitive and functional status: a prospective study. Can Med Assoc J 2001; 165: Murray AM, Levkoff SE, Wetle TT et al. Acute and functional decline in the hospitalized elderly patient. J Gerontol 1993; 48: M Inouye SK, Rushing JT, Foreman MD, Palmer RM, Pompei P. Does contribute to poor hospital outcomes? A threesite epidemiologic study. J Gen Intern Med 1998; 13: O Keeffe S, Lavan J. The prognostic significance of in older hospital patients. J Am Geriatr Soc 1997; 45: McCusker J, Cole MG, Dendukuri N, Belzile E. Does increase hospital stay? J Am Geriatr Soc 2003; 51: Francis J, Martin D, Kapoor WN. A prospective study of in hospitalized elderly. JAMA 1990; 263: Rockwood K, Cosway S, Carver D, Jarrett P, Stadnyk K, Fisk J. The risk of dementia and death after. Age Ageing 1999; 28: McCusker J, Cole MG, Abrahamowicz M, Primeau F, Belzile E. Delirium predicts 12-month mortality. Arch Intern Med 2002; 162: McCusker J, Cole MG, Dendukuri N, Han L, Belzile E. The course of in older medical inpatients: a prospective study. J Gen Intern Med 2003; 18: Marcantonio ER, Kiely DK, Simon SE et al. Outcomes of older people admitted to postacute facilities with. J Am Geriatr Soc 2005; 53: Oxman AD, Cook DJ, Guyatt GH. Users guides to the medical literature: VI. How to use an overview. JAMA 1994; 272: Laupacis A, Wells G, Richardson S, Tugwell P. Users guides to the medical literature: V. How to use an article about prognosis. JAMA 1994; 272: Egger M, Davey Smith G, Altman DG. Systematic Reviews in Health Care. Meta-Analysis in Context. London: BMJ Publishing Group, Hodkinson HM. Mental impairment in the elderly. J R Coll Physicians London 1973; 7: Rockwood K. Acute confusion in elderly medical patients. J Am Geriatr Soc 1989; 37: Levkoff SE, Evans DA, Liptzin B et al. Delirium: the occurrence and persistence of symptoms among elderly hospitalized patients. Arch Intern Med 1992; 152: Williams-Russo P, Urquhart BL, Sharrock NE, Charlson ME. Post-operative : predictors and prognosis in elderly orthopedic patients. J Am Geriatr Soc 1992; 40: Jitapunkul S, Pillay I, Ebrahim S. Delirium in newly admitted elderly patients: a prospective study. Quart J Med 1992; 300: Rockwood K. The occurrence and duration of symptoms in elderly patients with. J Gerontol 1993; 48: M Gaudet M, Pfitzenmeyer P, Tavenier-Vidal B, Lechenet M. Les états confusionnels en milieu interniste gériatrique court séjour. Psychologie Medicale 1993; 25: Rudberg MA, Pompei P, Foreman MD, Ross RE, Cassel CK. The natural history of in older hospitalized patients: a syndrome of heterogeneity. Age Ageing 1997; 26: Treloar AJ, Macdonald AJD. Outcome of. Part 1: outcome of diagnosed by DSM-III-R, ICD-10 and CAMDEX and derivation of the reversible cognitive dysfunction scale among acute geriatric inpatients. Int J Geriatr Psychiatry 1997; 12: Marcantonio ER, Flacker JM, Michaels M, Resnick NM. Delirium is independently associated with poor functional recovery after hip fracture. J Am Geriatr Soc 2000; 48: Kelly KG, Zisselman M, Cutillo-Schmitter T, Reichard R, Payne D, Denman SJ. Severity and course of in medically hospitalized nursing facility residents. Am J Geriatr Psychiatry 2001; 9: Kiely DK, Bergmann MA, Jones RN, Murphy KM, Orav JE, Marcantonio ER. Characteristics associated with persistence among newly admitted post-acute facility patients. J Gerontol Med Sci 2004; 59A: Lundstrom M, Edlund A, Karlsson S, Brannstrom B, Bucht G, Gustafson Y. A multifactorial intervention program reduces the duration of, length of hospitalization, and mortality in delirious patients. J Am Geriatr Soc 2005; 53: McAvay GJ, Van Ness PH, Bogardus ST et al. Older adults discharged from the hospital with : 1-year outcomes. J Am Geriatr Soc 2006; 54: Pitkala KH, Laurila JV, Strandberg TE, Tilvis RS. Multicomponent geriatric intervention for elderly inpatients with : a randomized, controlled trial. J Gerontol Med Sci 2006; 61A: Inouye SK, Zhang Y, Jones RN, Kiely DK, Yang F, Marcantonio ER. Risk factors for at discharge: development and validation of a predictive model. Arch Intern Med 2007; 167: Cole M, McCusker J, Ciampi A, Belzile E. The 6 and 12 month outcomes of older medical inpatients who recover from subsyndromal. J Am Geriatr Soc 2008; 56: Cole MG, You Y, McCusker J, Ciampi A, Belzile E. The 6 and 12 month outcomes of older medical inpatients who recover from. Int J Geriatr Psychiatry 2008; 23: Cole MG, Primeau FJ, Bailey RF et al. Systematic intervention for elderly inpatients with : a randomized trial. Can Med Assoc J 1994; 151: Cole MG, McCusker J, Bellavance F et al. Systematic detection and multidisciplinary care of in older medical patients: a randomized trial. Can Med Assoc J 2002; 167: Élie LM, Cole MG, Primeau FJ, Bellavance F. Delirium risk factors in elderly hospitalized patients. J Gen Intern Med 1998; 13: Cole MG, McCusker J, Dendukuri N, Han L. Symptoms of among elderly medical inpatients with or without dementia. J Neuropsychiatry Clin Neurosci 2002; 14: Blass JP, Gibson GE. Cerebrometabolic aspects of in relationship to dementia. Dement Geriatr Cogn Disord 1999; 10: Freeman GB, Mykytyn V, Gibson GE. Differential alteration of dopamine, acetylcholine, and glutamate release during anoxia and/or 3,4-diaminopyridine treatment. Neurochem Res 1987; 12: Adamis D, Treloar A, Martin FC, MacDonald AJ. A brief review of the history of as a mental disorder. Hist Psychiatry 2007; 18: Received 16 June 2008; accepted in revised form 29 August