Subsyndromal delirium in the ICU: evidence for a disease spectrum

Intensive Care Med (2007) 33:1007 1013 DOI 10.1007/s00134-007-0618-y ORIGINAL Sébastien Ouimet Richard Riker Nicolas Bergeon Mariève Cossette Brian Kavanagh Yoanna Skrobik Subsyndromal delirium in the ICU: evidence for a disease spectrum Received: 27 April 2006 Accepted: 5 March 2007 Published online: 3 April 2007 Springer-Verlag 2007 This study was partly funded by the FRSQ (Fonds de Recherche en Santé du Québec, Réseau respiratoire). Electronic supplementary material The online version of this article (doi:10.1007/s00134-007-0618-y) contains supplementary material, which is available to authorized users. S. Ouimet Y. Skrobik ( ) Maisonneuve-Rosemont Hospital and Université de Montréal, Intensive Care Unit, 5415 Boulevard del Assomption, H1T 2M4 Montreal, Canada e-mail: skrobik@sympatico.ca Tel.: +1-514-2523400; +1-514-2526229 Fax: +1-514-9398891 R. Riker Maine Medical Center, Division of Pulmonary/Critical Care Medicine, Portland Me., USA N. Bergeon Centre Hospitalier del Université de Montréal, Service of Consultation-Liaison Psychiatry, Montreal, Canada M. Cossette Montreal Heart Institute Coordinating Center, Biostatistics Service, 2600 William-Tremblay Street, H1Y 3J2 Montreal, Canada B. Kavanagh Hospital for Sick Children and University of Toronto, Departments of Anesthesia and Critical Care Medicine, Toronto, Canada Abstract Objective: ICU delirium is common and adverse. The Intensive Care Delirium Screening Checklist (ICDSC) score ranges from 0 to 8, with a score of 4 or higher indicating clinical delirium. We investigated whether lower (subsyndromal) values affect outcome. Patients: 600 patients were evaluated with the ICDSC every 8 h. Measurements and results: Of 558 assessed patients 537 noncomatose patients were divided into three groups: no delirium (score = 0; n = 169, 31.5%), subsyndromal delirium (score = 1 3; n = 179, 33.3%), and clinical delirium (score 4; n = 189, 35.2%). ICU mortality rates were 2.4%, 10.6%, and 15.9% in these three groups, respectively. Post-ICU mortality was significantly greater in the clinical delirium vs. no delirium groups (hazard ratio = 1.67) after adjusting for age, APACHE II score, and medication-induced coma. Relative ICU length of stay was: no delirium < subsyndromal delirium < clinical delirium and hospital LOS: no delirium < subsyndromal delirium clinical delirium. Patients with no delirium were more likely to be discharged home and less likely to need convalescence or long-term care than those with subsyndromal delirium or clinical delirium. ICDSC score increments higher than 4/8 were not associated with a change in mortality or LOS. Conclusions: Clinical delirium is common, important and adverse in the critically ill. A graded diagnostic scale permits detection of a category of subsyndromal delirium which occurs in many ICU patients, and which is associated with adverse outcome. Keywords Delirium Intensive care Screening Outcome Critical care Cognitive abnormalities Introduction Delirium in the critically ill is common [1, 2], morbid [3], and distressing [4]. These considerations as well as expert guidelines [5] have fostered initiatives [1 3] for reliable, easily applicable screening tools. Numerous tools are available to assess delirium in hospitalized patients outside the ICU [6 8]. Two scales tailored to mechanically ventilated patients have been validated to screen for delirium in the critically ill: the Intensive Care Delirium Screening Checklist (ICDSC) [9], and the Confusion Assessment Method-ICU [10]. Each has been used as

1008 dichotomous marker for delirium, i.e., they indicate that the patient either has delirium or does not. The ICDSC is an eight-item checklist producing a score between 0 and 8, where each item (corresponding to a value of 1 out of 8) reflects a Diagnostic and Statistical Manual of Mental Disorders IV (or alternative) diagnostic characteristic of delirium [9] (see Electronic Supplementary Material, ESM). The data are collected in real-time during each nursing shift. A ICDSC score of 4 or higher is well correlated with a formal psychiatric diagnosis of delirium (sensitivity 99%, specificity 64%, receiver operating characteristic curve area 0.901) [9] and has excellent interobserver reliability among nurses, as well as between nurses and critical care physicians [1, 9]. Outside the intensive care unit some elderly hospitalized patients never develop all the symptoms required for Diagnostic and Statistical Manual of Mental Disorders IV diagnosis of clinical delirium and may present with only some symptoms [11, 12]. Subsyndromal delirium occurs in patients with similar risks to those observed with delirium and is associated clinical outcomes that are intermediate between normal and delirious patients [11]. The ICDSC s potential advantage, because of its range, is that it may identify patients who, while falling short of the threshold diagnostic criteria for overt or clinical delirium, may nonetheless have clinically important subsyndromal illness. Moreover, when delirium is identified using this scale (i.e., any four or more symptoms are present out of a possible eight), the number of symptoms present at the initial diagnosis of delirium (or the change in that number over time) can be compared to outcome. We investigated whether patients with some features of delirium but not fulfilling diagnostic criteria-have associated adverse clinical outcomes. Methods All information was extracted anonymously from medical records for quality assurance purposes. Consent was waived by the institutional ethics committee and institutional review board at the Hôpital Maisonneuve Rosemont, which approved the study. Assessment of delirium All patients in the ICU for over 24 h were entered into the study. Moribund patients, and patients whose level of consciousness never permitted evaluation with the ICDSC throughout their ICU stay were excluded. Patients who developed coma of neurological cause (hemorrhage, stroke, or anoxia) while in the ICU were considered separately. Each patient s level of sedation was evaluated every shift using the Richmond Agitation and Sedation Scale (RASS), a diagnostic tool validated for ICU use [13], ranging from comatose (score of 5) to combative (score of +4). If the level of sedation permitted (i.e., RASS score between 3 and +4), an ICDSC evaluation was performed by the bedside nurse within 8 h of admission and daily at the end of each 8-h shift (see ESM) until discharge from the ICU. Changes in wakefulness and attention directly attributable to recent sedative medication were not scored as positive ICDSC points. Patients with a diagnostic ICDSC score (i.e., 4) at any assessment during their ICU stay were considered to have delirium [9]. Other data from this cohort were collected for purposes unrelated to the current study. Patients The study included 604 consecutive patients admitted to the ICU (January to December 2004) for more than 24 h. No patient was admitted to the ICU for delirium (alcohol-withdrawal related or other). 79% of patients were mechanically ventilated on admission, and 54% remained mechanically ventilated throughout their ICU stay (i.e., until the day of discharge or death). Mean patient RASS score was 0.4 ± 1.38. Three patients were excluded because they died within 36 h of admission, and data from one further patient were lost. Of the remaining 600 patients 42 had a level of consciousness that precluded adequate assessment for delirium, with RASS scores of 4 or 5 throughout their ICU stay. Complete assessments were available for the remaining 558 patients. Twenty-one patients developed coma of neurological cause (RASS score of 4 or 5) while in the ICU (18 deaths); these 21 were not included in the analysis. We attempted 3,309 assessments; 358 could not be completed because of transient RASS levels of 4 or 5. Incapacity to perform 244 additional assessments was associated with transient RASS levels of 3 (94%), 2 (5%), and agitation (RASS +2 or +3, < 1%). Mean ICDSC scores were calculated at least once each 8-h shift. The score during any shift during the ICU stay determined patients being categorized as follows: (a) no delirium, scored 0 every shift of the ICU stay (i.e., never manifested any of the eight items; n = 169); (b) subsyndromal delirium, scored 1 3 (i.e., manifested some of the eight items but never reached diagnostic criteria for clinical delirium; n = 179); (c) clinical delirium, scored 4 or higher (i.e., manifested at least four of the eight items; n = 189). No patient was either subsyndromally or clinically delirious prior to admission. Table 1 presents the demographic data for each patient group. Patients in the no delirium group were younger, had a higher proportion of surgical admission diagnoses and had the lowest Acute Physiology and Chronic Health Evaluation (APACHE) II scores at admission; the clinical delirium group had the highest admission APACHE II scores (Table 1).

1009 Table 1 Demographic characteristics of patients in the groups with no delirium (ND), subsyndromal delirium (SD), and clinical delirium (CD) All (n = 537) ND (n = 169) SD (n = 179) CD (n = 189) p a Age (years) 63 ± 15 60 ± 15 65 ± 14 64 ± 15 Overall 0.0062 ND vs. SD 0.0026 ND vs. CD 0.0138 SD vs. CD 0.5461 Men/women 59.6%/40.4% 60.4%/39.6% 55.9%/44.1% 62.4%/37.6% APACHE II 16.2± 7.9 12.9 ± 6.9 16.7 ± 7.8 18.6 ± 8.0 Overall < 0.0001 ND vs. SD < 0.0001 SD vs. CD 0.0162 Medical/surgical 47.5%/52.5% 37.9%/62.1% 52.5%/47.5% 51.3%/48.7% Overall 0.0101 ND vs. SD 0.0063 ND vs. CD 0.0110 SD vs. CD 0.8218 HTA 56.1% 52.1% 54.2% 61.4% ROH 13.0% 8.9% 10.6% 19.0% Overall 0.0085 ND vs. SD 0.5855 ND vs. CD 0.0071 SD vs. CD 0.0251 Drug-induced coma 18.1% 2.4% 16.8% 33.3% Overall < 0.0001 ND vs. SD 0.0001 SD vs. CD 0.0003 ICU LOS (days) 6.3± 8.2 2.5 ± 2.1 5.2 ± 4.9 10.8 ± 11.3 Overall < 0.0001 ND vs. SD < 0.0001 SD vs. CD < 0.0001 Hospital LOS (days) 36.4± 41.5 31.6 ± 46.5 40.9 ± 47.0 36.4 ± 28.9 Overall < 0.0001 ND vs. SD 0.0001 ND vs. CD 0.0001 SD vs. CD 0.9273 ICU mortality 9.9% 2.4% 10.6% 15.9% Overall < 0.0001 ND vs. SD 0.0020 SD vs. CD 0.1378 a Pairwise comparisons were performed only if the global F test was significant at the 0.05 level Recording of checklist scores The results of the checklist scoring were recorded as follows: (a) admission score, the initial score on the day of admission to the ICU; (b) maximum score, the highest score during the ICU stay; and (c) mean daily score, one score was obtained during each of three 8-h nursing shifts, and the average of these three scores constituted the mean daily score. Outcome Mortality and length of stay (LOS) in ICU and in hospital were recorded. ICU LOS was defined as the number of days spent in the ICU until transfer to the ward or death (if death occurred in the ICU). Hospital LOS was defined as the number of days from hospital admission until discharge from the hospital or death (if death occurred in the hospital). Mortality was based on status (alive vs. dead) at data collection closure, 31 May 2005, using inpatient and outpatient hospital records. Time to death was computed from the exit day of the first ICU admission (i.e., at the time when overall delirium status was determined) until death or censored at the last contact time, 31 May 2005. In addition, the degree of dependency following discharge from hospital was noted (i.e., home with no assistance, home with assistance, convalescence, chronic care, or other). Statistical analysis Demographic data are expressed as means ± SD for continuous variables and frequencies and percentages for categorical variables. Continuous variables were compared between the three groups using one-way analysis of variance models followed by pairwise comparisons if the global F test was significant at the 0.05 level. For continuous variables with skewed distribution the nonparametric Kruskal Wallis test were used, followed by the Mann Whitney test in the case of significant findings. Categorical variables were compared between

1010 the three groups using the χ 2 test. Mortality was studied using survival analysis. Time to death was computed from the date of patient exited the ICU for the first time. Fifty-three patients died during their first ICU stay (4 with no delirium, 19 with subsyndromal delirium, and 30 with delirium) and were not included in the survival analysis. Survival curves were computed using the Kaplan Meier formulas and compared between the three groups by the log rank test. In the case of significant findings 2 2 survival curves were also compared using the log rank test. Cox regression analysis was used to analyze the effect of group on mortality adjusted for age, severity of illness (APACHE II) and medication-induced coma [14]. Cox regression model was also performed on the clinical delirium group only to determine whether the initial score (4 8 ICDSC items) is a predictor of mortality. In this analysis time to death was computed from the date of patient manifested an ICDSC score of 4 or higher. All analyses were performed with SAS version 8.2 (SAS Institute, Cary, N.C., USA); differences at the level of p 0.05 were considered statistically significant. Mortality There were 53 ICU deaths: 2.4%, 10.6%, and 15.9% in the no delirium, subsyndromal delirium, and clinical delirium groups, respectively (no delirium vs. subsyndromal delirium, p = 0.0020; no delirium vs. clinical delirium, p < 0.0001; subsyndromal delirium vs. clinical delirium, p = 0.1378; Fig. 1). On the 484 surviving patients post- ICU mortality was higher in the clinical delirium than in the subsyndromal delirium and no delirium groups (clinical delirium vs. no delirium, log-rank test, p < 0.0001; clinical delirium vs. subsyndromal delirium, log-rank test, p = 0.0016 and subsyndromal delirium vs. no delirium, log-rank test, p = 0.2309; Fig. 2, Table 3). The results of the Cox regression analysis showed that only the clinical delirium group had a higher risk of mortality than the no Results Diagnostic categories Among the patients with subsyndromal delirium the number with scores of 3, 2, and 1 on the ICDSC were 67 (37.4%), 54 (30.2%), and 58 (32.4%), respectively. Of the 1,239 assessments completed among subsyndromal delirium patients 482 (39%) were scored 1, 368 (29.7%) 2, and 389 (31.3%) 3. In patients with clinical delirium the most frequent score was 4 (51.3%), with higher scores progressively less common (Table 2). Fig. 1 Post-ICU mortality. Kaplan Meier curves depicting time to death in the no delirium, subsyndromal delirium, and clinical delirium groups Table 2 Results of Cox regression for post-icu mortality analysis: time to death, computed from the date of patient first exited first ICU; 53 patients died in their first ICU (no delirium 4, subsyndromal delirium 19, clinical delirium 30) and were therefore not included in the analysis (HR hazard ratio, CI confidence interval) HR a 95% CI p Subsyndromal delirium Unadjusted model 1.32 0.84 2.08 0.2346 Adjusted model b 0.97 0.61 1.55 0.9098 Clinical delirium Unadjusted model 2.46 1.63 3.73 < 0.0001 Adjusted model b 1.67 1.07 2.60 0.0252 a Reference group is no delirium b Adjusted for age, APACHE II score, and coma Fig. 2 Distribution of highest checklist scores over time. All patients were considered nondelirious prior to admission. In patients with clinical delirium or subsyndromal delirium the maximal scores occurred early in the ICU admission, with almost one-half occurring within 24 48 h (i.e., within the first day) of admission, and over 90% occurring by ICU day 6. See Table 3

1011 Table 3 Mortality estimates from Kaplan Meier survival curves (percentages) in the groups with no delirium (ND), subsyndromal delirium (SD), and clinical delirium (CD). Post-ICU follow-up for the 484 remaining patients is 520 ± 299 days. The post-icu follow-up period varies from 1 to 874 days with a median at 634 days. After the 701st day, only censored patients remained (see Fig. 2) Days after ND (n = 165) SD (n = 160) CD (n = 159) ICU discharge 10 1.8 7.5 13.8 20 4.2 10.0 20.1 30 7.3 10.6 23.3 40 7.9 12.5 27.7 50 7.9 14.4 28.9 100 11.5 17.5 35.9 200 12.1 20.0 37.1 300 14.6 21.9 39.0 400 16.4 23.8 39.6 500 17.7 24.4 41.6 600 20.0 25.9 41.6 700 20.8 25.9 41.6 delirium group (hazard ratio 2.46, p < 0.0001), and this result remained statistically significant (hazard ratio 1.67, p = 0.0252) after adjusting for age, APACHE II score, and medication-induced coma [14] (Table 2). Incremental scores and outcome When only patients with clinical delirium were considered, the number of symptoms did not predict time to death (p = 0.9807, univariate Cox analysis; p = 0.5433 adjusted for APACHE II score, age, and presence of medicationinduced coma; Table 4). Those with an initial score of 4 had a subsequent mean daily score of 2.5 ± 1.1; those with 5had3.0± 1.2, those with 6 had 3.8 ± 1.1, those with 7 had 3.7 ± 2.2, and those with 8 had 3.1 ± 1.4 regardless of treatment. Since symptoms fluctuate, patients considered delirious thus had a mean daily score of 2.5 3.7 delirium features. The presence of four symptoms or more (regardless of which they were) was more important in determining outcome than the specific symptoms, or their combinations. Delirium scores over time The mean daily scores for the overall ICU stay were, by definition, 0.0 ± 0.0 in the no delirium group, 1.14 ± 0.75 in the subsyndromal delirium group, and 2.91 ± 1.30 in the clinical delirium group. The distribution of days following ICU admission on which a threshold score for subsyndromal delirium or clinical delirium was first recorded is illustrated in Fig. 2. Almost one-half of patients who developed either clinical delirium or subsyndromal delirium first achieved their defining score within 24 48 h of admission to ICU (ICU day 1 ended at 24.00 hours on the day of admission). Over 90% of subsyndromal group achieved the score by day 6 (Fig. 2). Patients manifesting delirium on various days did not differ in terms of their characteristics. The (unadjusted) ICU length of stay was least in the absence of delirium; the groups rank order was no delirium < subsyndromaldelirium < clinical delirium (p < 0.0001, all pair-wise comparisons; Table 1). This pattern was not observed for hospital length of stay, where the rank order was no delirium < subsyndromal delirium not statistically different from clinical delirium (no delirium vs. subsyndromal delirium, p = 0.0001; no delirium vs. clinical delirium, p = 0.0001; subsyndromal delirium vs. clinical delirium, p = 0.9273; Table 1). Patients status following discharge from hospital is summarized in Fig. 3. Patients with no delirium had a significantly greater probability of being discharged home without the need for additional assistance than those diagnosed with either subsyndromal or clinical delirium. In contrast, the need for convalescence or long-term care was greater in those with either subsyndromal or clinical delirium than in those with no delirium. Subsyndromal and clinical delirium patients did not differ in terms of need for convalescence or long-term care (no delirium vs. subsyndromal delirium, p < 0.0001; no delirium vs. Table 4 ICDSC score upon diagnosis, and outcome in patients with clinical delirium No. of No. of patients No. of deaths ICU LOS ICDSC n % n % (days) items Four 97 51.3 51 52.6 13.3 ± 13.7 Five 47 24.8 22 46.8 8.8 ± 8.3 Six 30 15.8 17 56.7 7.8 ± 6.3 Seven 11 5.8 3 27.3 6.5 ± 6.4 Eight 4 2.1 3 75.0 7.8 ± 5.2 Fig. 3 Outcome following hospital discharge. Patients with no delirium were more frequently discharged home unassisted than those with either subsyndromal or clinical delirium (* p = 0.0004). The need for convalescence or long-term care was also less in those with no delirium than in those with either subsyndromal delirium or clinical delirium (*p < 0.001)

1012 clinical delirium, p =0.0012; subsyndromal delirium vs. clinical delirium, p = 0.7519; Fig. 3). Risk factors Clinical features associated with delirium [14] were also associated with subsyndromal delirium: incremental age, APACHE II scores, and the presence of medicationinduced coma (Table 1). A history of hypertension did not differ significantly between the no delirium, subsyndromal delirium, and clinical delirium groups in this cohort (p = 0.17). Alcohol abuse differed only between patients with clinical delirium and the other two cohorts (p = 0.008 overall; 2 2 comparison between clinical delirium and subsyndromal delirium, p = 0.58). Discussion This study suggests that an entity of subsyndromal delirium exists in critically ill patients, and that it is associated with clinically important adverse outcome. Patients presenting with this syndrome fall into an intermediate category which differs from both no delirium and clinical delirium. Delirium in ICU: a spectrum ICU patients with one to three items on the ICDSC checklist have a poorer outcome than those with no items. Since the hazard ratio for mortality is not statistically significant when controlling for potential confounders, a subsyndromal delirium score may present a marker of disease severity rather than an independent risk factor in this population. Nonetheless, the associations (for subsyndromal vs. no delirium) remain significant for ICU length of stay, hospital length of stay, and status following discharge from hospital. Previous authors [11] have reported intermediate outcomes for elderly patients outside the ICU who have symptoms of delirium, but in whom the full clinical manifestations of delirium are never reached. The current data suggest a similar spectrum from no delirium through subsyndromal delirium to clinical delirium in the critically ill, and one that is associated with a parallel spectrum of adverse outcome. Some risk factors (age, APACHE II score, and the presence of medication induced-coma) associated with subsyndromal delirium resemble those described for delirium. Others (such as alcohol abuse) do not. The pathogenesis of delirium may therefore have common associated predisposing factors or mediators, and others which vary from those associated with subsyndromal delirium. The concept of a continuous instead of a categorical (or dichotomous) characterization of delirium may allow us to better understand the association of psychiatric symptoms with clinical outcomes. For example, Lesperance et al. [15] have reported a dose-dependent association between the level of depressive symptoms during admission for myocardial infarction and long-term cardiac mortality, independently of whether the diagnostic threshold for major depression was reached. Recognition of a spectrum of incidence and of effect may also provide a partial explanation for the widely discrepant reported incidences of delirium in ICU. Various authors have reported incidences ranging from 11% [16] to over 80% [3]. Screening and diagnostic methods differ among such studies and may include patient assessments performed at different times during the ICU day. Only 30.2% of our cohort never manifested any of the ICDSC items; these patients would likely have been considered cognitively normal by most assessors. This means that 70% of our patients did at some time during their ICU stay demonstrate at least one feature of delirium. Despite not fulfilling psychiatric criteria for clinical delirium, it is possible that application of a sensitive tool would detect an incidence of delirium of up to 70% in a population with characteristics similar to those of the current cohort. On the basis of ICDSC scores we considered 36.5% of these patients to have clinical delirium. Although this incidence is somewhat higher than the 19% incidence we previously reported [9], the higher APACHE II scores may account for the higher incidence of delirium. In the current study patients with subsyndromal delirium (scores 1 3) constituted 33% of the cohort. Such patients could have a pivotal effect in dichotomous characterizations of delirium because their inclusion among patients considered not to have delirium (i.e., normal patients) would worsen that group s prognosis. Conversely, if such patients were included among those considered to have delirium, the resulting group s prognosis would be improved. We found these subsyndromal patients more likely to survive but just as likely to have longer hospitals stays or require institutionalization at discharge as patients with clinical delirium. Threshold effects Patients with clinical delirium have higher mortality and greater LOS regardless of score than those with no delirium and thus have the worst prognosis of all. The current results suggest that the ICDSC offers a useful checklist for screening patients with few or no features of clinical delirium. However, when delirium is present, the checklist does not differentiate between moderate and severe delirium. Thus the ICDSC score may simply indicate that across all ranges of symptom frequency delirium contributes to outcome, but that a plateau exists at the threshold of clinical

1013 delirium beyond which additional symptoms do not contribute further to adverse outcome. If true, this would correspond to the maximal CNS burden of disease that is translatable into patient outcome. Conclusion The eight items of the ICDSC allow graded symptom identification in patients, and such identification is associated with prognosis. As with the hospitalized elderly, checklist scores indicating intermediate levels of symptoms appear to identify a group ( subsyndromal delirium ) at significant risk of prolonged ICU and hospital LOS and with greater dependence following hospital discharge. These risks are intermediate between those associated with the absence of delirium (lowest risk) and those associated with the presence of clinical delirium (highest risk). Identification of patients with such subsyndromal delirium may allow prevention or treatment in patients previously unsuspected as being at higher risk, which may in turn translate into improved outcome. References 1. Dubois MJ, Bergeron N, Dumont M, Dial S, Skrobik Y (2001) Delirium in an intensive care unit: a study of risk factors. Intensive Care Med 27:1297 1304 2. Lin SM, Liu CY, Wang CH, Lin HC, Huang CD, Huang PY, Fang YF, Shieh MH, Kuo HP (2004) The impact of delirium on the survival of mechanically ventilated patients. Crit Care Med 32:2254 2259 3. Ely EW, Shintani A, Truman B, Speroff T, Gordon SM, Harrell FE Jr, Inouye SK, Bernard GR, Dittus RS (2004) Delirium as a predictor of mortality in mechanically ventilated patients in the intensive care unit. JAMA 291:1753 1762 4. Breitbart W, Gibson C, Tremblay A (2002) The delirium experience: delirium recall and delirium-related distress in hospitalized patients with cancer, their spouses/caregivers, and their nurses. Psychosomatics 43:183 194 5. Jacobi J, Fraser GL, Coursin DB, Riker RR, Fontaine D, Wittbrodt ET, Chalfin DB, Masica MF, Bjerke HS, Coplin WM, Crippen DW, Fuchs BD, Kelleher RM, Marik PE, Nasraway SA Jr, Murray MJ, Peruzzi WT, Lumb PD (2002) Clinical practice guidelines for the sustained use of sedatives and analgesics in the critically ill adult. Crit Care Med 30:119 141 6. Breitbart W, Rosenfeld B, Roth A, Smith MJ, Cohen K, Passik S (1997) The Memorial Delirium Assessment Scale. J Pain Symptom Manage 13:128 137 7. Inouye SK, van Dyck CH, Alessi CA, Balkin S, Siegal AP, Horwitz RI (1990) Clarifying confusion: the confusion assessment method. A new method for detection of delirium. Ann Intern Med 113:941 948 8. O Keeffe ST, Mulkerrin EC, Nayeem K, Varughese M, Pillay I (2005) Use of serial Mini-Mental State Examinations to diagnose and monitor delirium in elderly hospital patients. J Am Geriatr Soc 53:867 870 9. Bergeron N, Dubois MJ, Dumont M, Dial S, Skrobik Y (2001) Intensive Care Delirium Screening Checklist: evaluation of a new screening tool. Intensive Care Med 27:859 864 10. Ely EW, Inouye SK, Bernard GR, Gordon S, Francis J, May L, Truman B, Speroff T, Gautam S, Margolin R, Hart RP, Dittus R (2001) Delirium in mechanically ventilated patients: validity and reliability of the confusion assessment method for the intensive care unit (CAM-ICU). JAMA 286:2703 2710 11. Cole M, McCusker J, Dendukuri N, Han L (2003) The prognostic significance of subsyndromal delirium in elderly medical inpatients. J Am Geriatr Soc 51:754 760 12. Levkoff SE, Besdine RW, Wetle T (1986) Acute confusional states (delirium) in the hospitalized elderly. Annu Rev Gerontol Geriatr 6:1 26 13. Ely EW, Truman B, Shintani A, Thomason JW, Wheeler AP, Gordon S, Francis J, Speroff T, Gautam S, Margolin R, Sessler CN, Dittus RS, Bernard GR (2003) Monitoring sedation status over time in ICU patients: reliability and validity of the Richmond Agitation-Sedation Scale (RASS). JAMA 289:2983 2991 14. Ouimet S, Kavanagh BP, Gottfried SB, Skrobik Y (2006) Incidence, risk factors and consequences of ICU delirium. Intensive Care Med 33:66 73 15. Lesperance F, Frasure-Smith N, Talajic M, Bourassa MG (2002) Five-year risk of cardiac mortality in relation to initial severity and one-year changes in depression symptoms after myocardial infarction. Circulation 105:1049 1053 16. Kishi Y, Iwasaki Y, Takezawa K, Kurosawa H, Endo S (1995) Delirium in critical care unit patients admitted through an emergency room. Gen Hosp Psychiatry 17:371 379