Effect of Sedation Level on the Prevalence of Delirium When Assessed with CAM-ICU and ICDSC

Electronic Supplementary Material:

Effect of Sedation Level on the Prevalence of Delirium when Assessed with CAM-ICU and ICDSC

Matthias Haenggi, Sina Blum*, Ruth Brechbuehl*, Anna Brunello, Stephan M. Jakob, Jukka Takala

Sedation concept:
In our ICU the nurse-implemented sedation protocol mandates sedative boli, and allows continuous sedation only if more than 6 boli per 4 hours are needed. The use of propofol is encouraged; midazolam is reserved for hemodynamically unstable patients. Pain medication consists of scheduled intravenous acetaminophen and additional fentanyl boli as needed. Continuous fentanyl is rarely used.

Delirium and sedation level assessment:
Delirium assessment with CAM-ICU and ICDSC was established in this ICU prior to this study [1], but preceding study start, all nursing staff, fellows, residents and attending physicians received a 50-minute session on delirium assessment – a refresher of CAM-ICU, and introduction of ICDSC, including material downloaded from the Vanderbilt University Medical Center’s webpage We did not formally test the success of the introduction of ICDSC and the impact of the refresher training of CAM-ICU.

Presence or absence of delirium was evaluated using CAM-ICU as initially developed and presented by Ely et al. [2], in a German or French translation, provided at CAM-ICU considers patients delirious when an acute onset of altered mental status or its fluctuation is accompanied by inattention and either an altered level of consciousness or disorganized thinking. CAM-ICU was assessed together with RASS.

Delirium was also assessed by a second scale, the ICDSC, developed by Bergeron et al. [3] in a German translation [4]. The domains like inattention, disorganized thinking and level of consciousness were evaluated during sedation stop. The evaluation of other domains required observations during a whole shift (e.g. psychomotor activity, speech, mood and sleep disturbance or fluctuation of symptoms), therefore additional information was obtained from the bedside nurse in charge of the patient. The ICDSC considers patients are delirious when at least four of the above 8 items are deviant.

The level of consciousness was determined with the Richmond Agitation-Sedation Scale (RASS), a scale ranging from –5 (unarousable) to +4 (combative)[5].

When the sedation stop was initiated, the assessors (clinical team) followed the patient up until he/she had reached a stable level of RASS (at least -3 allowing delirium assessment) without further changes, for a maximum of 120 minutes after stopping the drugs. In some instances, sedation stop had to be discontinued prematurely due to clinical reasons, typically hemodynamic or gas exchange instability, or poor tolerance of mechanical ventilation or endotracheal tube, and sedation was resumed. If the patient remained comatose (RASS level -4 to -5), the observation was stopped after 2 hours and the patient was not assessed for delirium due to coma.

Statistical Details:

Because in clinical practice either CAM-ICU or ICDSC is performed, the analysis were stratified and done once for CAM-ICU and once for ICDSC. Post-hoc, as a second step, we included in the models the medication each patient received, and their interaction with the assessment method, to confirm that the differences between assessment methods are not strongly influenced by these. The interactions were removed from the model if insignificant (p > 0.05), and probabilities were calculated under the different assessment methods, while controlling for medication used. The effects of the different medications are reported in the electronic supplemented material. In addition, the analyses were repeated excluding patients with neurological admissions.

Differences in incidence was analysed with the z-test (chi-square-test within one category), and applying the Yates Correction Factor.

Reference List electronic supplementary material

1.Ruokonen E, Parviainen I, Jakob SM, Nunes S, Kaukonen M, Shepherd ST, Sarapohja T, Bratty JR, Takala J, Dexmedetomidine for Continuous Sedation I, (2009) Dexmedetomidine versus propofol/midazolam for long-term sedation during mechanical ventilation. Intensive Care Med 35: 282-290 DOI 10.1007/s00134-008-1296-0

2.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 DOI 11730446

3.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 DOI 11430542

4.Radtke FM, Franck M, Oppermann S, Lutz A, Seeling M, Heymann A, Kleinwachter R, Kork F, Skrobik Y, Spies CD, (2009) [The Intensive Care Delirium Screening Checklist (ICDSC)--translation and validation of intensive care delirium checklist in accordance with guidelines]. Anasthesiologie, Intensivmedizin, Notfallmedizin, Schmerztherapie : AINS 44: 80-86 DOI 10.1055/s-0029-1202647

5.Sessler CN, Gosnell MS, Grap MJ, Brophy GM, O'Neal PV, Keane KA, Tesoro EP, Elswick RK, (2002) The Richmond Agitation-Sedation Scale: validity and reliability in adult intensive care unit patients. Am J Respir Crit Care Med 166: 1338-1344 DOI 10.1164/rccm.2107138

Table 1. Type of sedation prior to scheduled sedation stop. Individual patients could receive more than one type of sedation during the study – hence the sum of patients receiving each type of sedation exceeds the total number of patients in the study. The last column shows the proportion of the use of propofol and midazolam before assessment.

Patients n
(% of total 80) / assessments n
(% of total 467) / number of assessments per patient during study / propofol vs midazolam (%)
Continuous sedation / 54 (67.5%) / 180 (38.5%) / 2.3 ± 3.1 / 80 vs 20
Intermittent sedation / 52 (65%) / 153 (32.8%) / 1.9 ± 2.8 / 68 vs 32
No sedation, intubated / 39 (48.8%) / 101 (21.6%) / 1.3 ± 2.0
No sedation, not intubated / 18 (22.5%) / 33 (7.1%) / 0.4 ± 1.1

Table 2 Prevalence of CAM-ICU/ICDSC positive assessments after exclusion of 32 patients with neurological admission diagnosis

CAM-ICU pos
original / CAM-ICU pos
without
RASS -3 and -2 / CAM-ICU pos
without
RASS -3 and -2
“not vigilant”
prevalence / 107/257
41.6% / 48/257
18.7 % / 85/257
33.1 %
Probability of being
classified positive / 46.9% / 22.4%* / 32.9%*
ICDSC pos
original / ICDSC pos
without
RASS -3 and -2 / ICDSC pos
without
RASS -3 and -2
“not vigilant”
prevalence / 113/253
44.7 % / 53/253
20.9 % / 89/253
35.2 %
Probability of being
classified positive / 43.1% / 20.7%* / 28.2%*

*= p-value < 0.001;

Table 3: Admission diagnosis, demographics and type of sedation before sedation stop of the 11 patients who were CAM-ICU positive only at RASS-2/-3. All patients received intermittent fentanyl as analgesic.

Patient ID / Age (years) / Admission diagnosis / SAPS II score / Number of assessments / Type of sedation
9 / 43 / Epilepsia / 72 / 1 / 1x propofol continuous
11 / 62 / pneumonia / 52 / 2 / 1x midazolam cont.
1x propofol intermittent
18 / 52 / Respiratory failure due to Guillain-Barré-Syndrome / 64 / 1 / 1x propofol cont.
29 / 49 / traumatic brain injury / 64 / 5 / 2x propofol cont.
3x propofol interm.
36 / 63 / Sepsis due to endocarditis / 50 / 1 / 1x propofol cont.
51 / 63 / pneumonia / 57 / 2 / 2x midazolam intermittent
55 / 71 / complicated cardiac surgery / 77 / 1 / 1x propofol interm.
72 / 53 / traumatic brain injury / 41 / 9 / 4x propofol cont.
1x propofol interm.
2x fentanyl only
2x none
73 / 59 / respiratory failure following traumatic brain injury / 48 / 1 / 1x propofol cont.
76 / 83 / traumatic brain injury / 50 / 1 / 1x fentanyl only
83 / 49 / pneumonia / 61 / 2 / 2x midazolam intermittent
average of all patients / 61 ± 17 / 40% admission for neuro/neurotrauma / 55 ± 18 / 5.9 ± 5.1

Table 4: numbers of patients with the different types of delirium, diagnosed according either CAM-ICU or ICDSC

Coma only / Delirium only at RASS-2/-3
(hypoactive) / Hypoactive delirium / Hyperactive delirium / Mixed / No delirium
CAM-ICU / 9 / 11 / 20 / 8 / 19 / 13
ICDSC / 9 / 9 / 18 / 8 / 19 / 17

Table 5 ANOVA on ranks (p-values) of length of stay (LOS) in the ICU or in the hospital and probability of death/logistic regression (p-values) 28-day or 90-day mortality

ICU LOS / Hospital LOS / 28-day mortality / 90-day mortality
Pat ever CAM-ICU positive
vs. never positive / 0.15 / 0.57 / 0.39 / 0.45
Pat ever CAM-ICU pos (excluding RASS -2/-3) vs never positive / 0.10 / 0.38 / 0.70 / 0.96
Pat ever CAM-ICU pos (excluding insufficient vigilance at RASS -2/-3)
vs. never positive / 0.05 / 0.58 / 0.41 / 0.54
Pat ever ICDSC pos
vs. never positive / 0.01 / 0.61 / 0.78 / 0.92
Pat ever ICDSC pos (excluding RASS -2/-3) vs. never positive / 0.01 / 0.78 / 1 / 0.69
Pat ever ICDSC pos (excluding insufficient vigilance at RASS -2/-3)
vs. never positive / 0.00 / 0.90 / 0.69 / 0.88

Table 6: admission data, with details of diagnosis and preexisting illness

Admission diagnosis / Pre-exixsting neurologic disorders / Admitted for neuro/neuro-trauma / age / SAPS II
Intracerebral hemorrhage / no / yes / 60 / 47
ARDS after post laparotomy / no / no / 54 / 58
Anaphylaxis / no / no / 79 / 27
Multiple trauma including traumatic brain injury (tbi) / no / yes / 62 / 61
Post cardiac surgery / no / no / 84 / 50
Post cardiac surgery / no / no / 31 / 69
Tbi / no / yes / 74 / 26
Status epilepticus / no / yes / 62 / 72
Spinal trauma / no / yes / 73 / 32
Pneumonia / no / no / 62 / 52
Post cardiac surgery / no / no / 63 / 39
Subarachnoidal hemorrhage / no / yes / 75 / 58
Multiple trauma including tbi / no / yes / 25 / 44
ARDS post lung surgery / Focal epilepsia (treated) / no / 82 / 59
Typ B aortic dissection / narcolepsia / no / 71 / 36
Guillan Barré Syndrom / no / yes / 70 / 64
Stroke / no / yes / 67 / 32
Liver failure / Cerebral lymphoma, HIV / yes / 45 / 86
Tbi / no / yes / 34 / 65
Mediastinitis after epiglotitis / no / no / 78 / 34
Sepsis post CABG / no / no / 60 / 63
Infected aortic graft, mycotic aneurysm/stroke / no / yes / 57 / 86
Tbi / no / yes / 61 / 41
Typ A aortic dissection / no / no / 65 / 70
Tbi / Post intracrebral hemorrhage / yes / 75 / 64
Multiple trauma incl. tbi / no / yes / 61 / 27
Urosepsis / no / no / 67 / 79
Heart failure / no / no / 69 / 63
Post cardiac surgery / Post transitory ischemic attack / no / 53 / 54
Typ A aortic dissection / no / no / 59 / 57
Sepsis, susp. meningitis / no / yes / 46 / 84
Endocarditis with cerebral embolies / no / yes / 70 / 50
Intracerebral hemorrhage / no / yes / 83 / 49
Tbi / Multiple sclerosis / yes / 80 / 26
COPD / no / no / 75 / 39
Post lapartomie / no / no / 52 / 57
Tbi / no / yes / 31 / 43
Tbi / no / yes / 20 / 65
Tbi / no / yes / 61 / 23
Sepsis / no / no / 49 / 67
Post cardiac surgery / no / no / 59 / 89
Sepsis / no / no / 57 / 44
Post cardiac surgery / no / no / 22 / 89
Pneumonia / no / no / 78 / 57
Post cardiac surgery / no / no / 80 / 54
acute liver an renal failure / no / no / 69 / 75
Pneumonia / no / no / 45 / 76
Post cardiac surgery / Parkinson’s disease / no / 60 / 77
Meningitis / no / yes / 54
Sepsis, suspicion meningitis / no / yes / 79 / 42
Stroke / no / yes / 62 / 35
Intracerebral hemorrhage / no / yes / 84 / 27
Sepsis / no / no / 31 / 61
Post thoracic surgery / no / no / 74 / 80
Post cardiac surgery / no / no / 62 / 75
Sepsis / no / no / 73 / 40
Post cardiac surgery / no / no / 62 / 59
Cardiogenic shock / no / no / 63 / 36
Respiratory failure post cardiac surgery / no / no / 75 / 45
Subarachnoidal hemorrhage / no / yes / 25 / 56
Post cardiac surgery / no / no / 82 / 50
Tbi / no / yes / 71 / 41
ARDS/pulmonary hypertension / Post tbi / no / 70 / 48
Hepatorenal Syndrome / no / no / 67 / 101
Decompensated cor pulmonale post laparotomy / no / no / 45 / 33
Tbi / no / yes / 34 / 50
Heart failure / no / no / 78 / 52
Hodgkin Lymphoma, acute coronary syndrome / no / no / 60 / 62
Heart failure / no / no / 57 / 60
Drug overdose (mixed) with epilepsia / Substance abuse with epilepsia / yes / 61 / 53
Multiple trauma incl tbi / no / yes / 65 / 54
Stroke / no / yes / 75 / 57
Pneumonia / no / no / 61 / 61
Sepsis / no / no / 67 / 62
Heart failure / no / no / 69 / 79
Multiple trauma / no / no / 53 / 32
ARDS / no / no / 59 / 75
Post cardiac surgery / Probable dementia (mini-mental score 28/30) / no / 46 / 17
Acute coronary syndrome / no / no / 70 / 42
Pneumonia / no / no / 83 / 71

Table 7: Results of the secondary analysis 1: probabilities of being classified as delirious, mixed effects logistic regression while controlling for medication given (all p<0.001).

CAM-ICU / CAM-ICU,
after excluding 32 patients with neurologic admission diagnosis / ICDSC / ICDSC,
after excluding 32 patients with neurologic admission diagnosis
original / 47.0% / 46.1% / 42.0% / 42.0%
without
RASS -3 and -2 / 20.1% / 20.5% / 19.5% / 19.2%
without
RASS -3 and -2
“not vigilant” / 29.9% / 31.1% / 27.2% / 25.9%

Table 8: Results of the secondary analysis 2: mixed effects logistic regression including medication as a co-factor. The reported results are the marginal probabilities under baseline conditions, which are the original assessment method (CAM-ICU and ICDSC, not the “without RASS-2 to -3” and the “without RASS -3 and -2 not vigilant”) and no medication. Under other conditions the probability would be slightly different. However, the differences between the probabilities will remain the same since the interaction assess-method/medication is not significant. The results should be interpreted cautiously because of collinearity problems.

CAM-ICU / CAM-ICU,
after excluding 32 patients with neurologic admission diagnosis / ICDSC / ICDSC,
after excluding 32 patients with neurologic admission diagnosis
no medication / 47.0% / 46.1% / 42.0% / 42.0%
propofol intermittent / 43.8% / 42.8% / 56.9% / 48.9%
midazolam intermittent / 37.0% / 41.4% / 25.4% / 25.2%
propofol continuous / 50.4% / 52.7% / 44.4% / 47.5%
midazolam intermittent / 16.7% / 16.8% / 13.3% / 10.7%
other / 99.8% / 100.0% / 99.5% / 100.0%