Delirium in critically ill patients: Impact on long-term health-related quality of life and cognitive functioning
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
Delirium in critically ill patients: Impact on long-term
health-related quality of life and cognitive functioning*
Mark van den Boogaard, RN, MSc; Lisette Schoonhoven, RN, PhD; Andrea W.M. Evers, PhD;
Johannes G. van der Hoeven, MD, PhD; Theo van Achterberg, RN, PhD; Peter Pickkers, MD, PhD
Objective: To examine the impact of delirium during intensive found between delirium and nondelirium survivors on the Short
care unit stay on long-term health-related quality of life and Form-36 and checklist individual strength-fatigue. However, sur-
cognitive function in intensive care unit survivors. vivors who had suffered from delirium reported that they made
Design: Prospective 18-month follow-up study. significantly more social blunders, and their total cognitive failure
Setting: Four intensive care units of a university hospital. questionnaire score was significantly higher, compared to survi-
Patients: A median of 18 months after intensive care dis- vors who had not been delirious. Survivors of a hypoactive delir-
charge, questionnaires were sent to 1,292 intensive care survivors ium subtype performed significantly better on the domain mental
with (n ⫽ 272) and without (n ⫽ 1020) delirium during their health than mixed and hyperactive delirium patients. Duration of
intensive care stay. delirium was significantly correlated to problems with memory
Measurements and Main Results: The Short Form-36v1, check- and names.
list individual strength-fatigue, and cognitive failure question- Conclusions: Intensive care survivors with delirium during
naire were used. Covariance analysis was performed to adjust for their intensive care unit stay had a similar adjusted health-related
relevant covariates. Of the 915 responders, 171 patients were quality of life evaluation, but significantly more cognitive prob-
delirious during their intensive care stay (median age 65 [inter- lems than those who did not suffer from delirium, even after
quartile range 58 – 85], Acute Physiology and Chronic Health Eval- adjusting for relevant covariates. In addition, the duration of
uation II score 17 [interquartile range 14 –20]), and 745 patients delirium was related to long-term cognitive problems. (Crit Care
were not (median age 65 [interquartile range 57–72], Acute Phys- Med 2012; 40:112–118)
iology and Chronic Health Evaluation II score 13 [interquartile KEY WORDS: cognition; critical care; delirium; health-related
range 10 –16]). After adjusting for covariates, no differences were quality of life; intensive care
D elirium is a disorder that fre- and long-term cognitive impairment (5, ium subtypes were performed (6, 13). A
quently occurs in intensive 6). Generally, without distinguishing be- significant difference between delirium
care unit (ICU) patients (1– tween delirium and nondelirium pa- and nondelirium patients in role-
3), and is recognized as acute tients, 25% to 78% of ICU patients expe- physical function, which mostly reflects
brain dysfunction with changes in con- rience cognitive impairments after functioning in daily activities, was re-
sciousness and cognition, which fluctu- discharge from the ICU (7), emphasizing ported (13), however, no correction for
ate during the day (4). This disorder is the need for more attention in the period disease severity was performed (13).
associated with serious health problems following critical illness. There is a grow- This implies that these findings could
ing interest in health-related quality of be the result of an epiphenomenon. The
life (HRQoL) after ICU discharge (8 –13). duration of delirium during patients’
*See also p. 319.
From the Department of Intensive Care Medicine
HRQoL questionnaires are usually subdi- ICU stay was associated with their ob-
(MvdB, JGvdH, PP), Scientific Institute for Quality of vided into dimensions relating to physi- served impaired cognitive performance
Healthcare (LS, TvA), Department of Medical Psychol- cal, mental, and social functioning. It is (6). Little is known about the long-term
ogy (AWE), and Nijmegen Institute for Infection, Inflam- recognized that the value of measure- (⬎1 yr) effects of delirium on aspects of
mation and Immunity (N4i) (JGvdH, PP), Radboud Uni-
versity Nijmegen Medical Centre, The Netherlands.
ments of cognitive functioning with a the HRQoL in this specific group of
Mr. van den Boogaard carried out the study, the general HRQoL questionnaire is limited patients. In addition, it is unknown if
statistics and drafted the manuscript. Ms. Schoon- in this setting, and specific surveys mea- there are differences in HRQoL (includ-
hoven and Dr. Pickkers were involved in the design, suring patients cognitive functioning, ing cognitive function) for subtypes of
supervised the study and helped to draft the manu-
script. Drs. Evers, van der Hoeven, and van Achterberg
such as the validated self-reporting cog- delirium (3) or if there is a correlation
supervised the study and corrected the manuscript. All nitive failure questionnaire (CFQ) (14), between the duration of delirium and
authors read and approved the final manuscript. have been developed. HRQoL.
The authors have not disclosed any potential con- Only two studies have examined the Therefore, the aim of this study was to
flicts of interest.
impact of delirium on HRQoL in ICU compare the HRQoL, including self-
For information regarding this article, E-mail:
m.vandenboogaard@ic.umcn.nl survivors (6, 13). These studies were reported cognitive functioning, in ICU
Copyright © 2012 by the Society of Critical Care rather small, relatively short with a survivors with delirium during their ICU
Medicine and Lippincott Williams & Wilkins maximum follow-up of 3 and 12 stay with those that did not suffer from
DOI: 10.1097/CCM.0b013e31822e9fc9 months, and no analyses of the delir- delirium, after a median of 18 months
112 Crit Care Med 2012 Vol. 40, No. 1after ICU discharge. Furthermore, we ex- delirium subtype of these patients were normally distributed, log transformation of all
amined the correlations between dura- collected. HRQoL data was carried out successfully and
tion of delirium and HRQoL, and if sub- At a median 18 months after ICU dis- the duration of delirium was divided into
types of delirium exerted different effects charge, an HRQoL survey was sent out to the quartiles, resulting in normally distributed
on HRQoL. cohort of ICU survivors. Four weeks after this, outcome measurements. The correlation be-
a reminder letter was sent to the nonre- tween duration of delirium divided into quar-
sponders. We used three different validated tiles and the log-transformed HRQoL was
MATERIALS AND METHODS instruments to measure the HRQoL. We will tested using Pearson’s correlation coefficient.
refer to these three tests as the HRQoL. Al- Significant differences in demographic vari-
though there is no specific HRQoL instrument ables between nondelirium and delirium pa-
Subjects
for ICU patients, recommended instruments tients and differences between the delirium
All consecutive patients admitted to the for ICU patients are the Short Form-36 (SF- subtypes were considered as covariates and a
ICU of the Radboud University Nijmegen Med- 36) and the EuroQoL-5D (20). We used the multivariate analysis of covariance was per-
ical Centre between February 2008 and Feb- validated Dutch version of the SF-36 version 1 formed. Since there was no difference in age
ruary 2009 were screened for delirium three (21), containing eight multi-item dimensions: between delirium and nondelirium responders
times a day with the confusion assessment physical functioning, role-physical, bodily in our population, adjusting for age was un-
method (CAM)-ICU (1, 15) by well-trained ICU pain, general health, vitality, social function- necessary. In view of the explorative nature of
nurses (16). In February 2010, after a median ing, role-emotional, and mental health. Aggre- this study, and to increase its sensitivity, no
duration of 18 months after ICU discharge, we gated summary scores were calculated for correction for multiple testing was performed.
evaluated the HRQoL of the surviving pa- physical and mental functioning, expressed in Statistical significance was defined as a p
tients. The regional Medical Ethical Commit- physical component score and mental compo- value ⬍ .05. All data were analyzed using
tee approved the study (study number 2010/ nent score, respectively. To calculate the phys- SPSS version 16.01 (SPSS, Chicago, IL).
008) and waived the need for informed ical component score and mental component
consent, since the objective of this study was score, we used the standardized Dutch popu-
RESULTS
to evaluate regular patient care. lation scores (22). In line with the SF-36
Health Survey Manual (23), missing values At the median of 18 months before
were imputed, and data were recoded and sub- this HRQoL survey, a total of 1,613 con-
Procedures sequently scored (range 0 to 100). A higher secutive patients who fulfilled the inclu-
score indicates a higher level of functioning. sion criteria were admitted (Fig. 1). In
All ICU patients were included in this study Additionally, the shortlist of the Dutch vali- this group, 1,202 patients had no delir-
except those who: were admitted for ⬍1 day; dated checklist individual strength (CIS)-
ium and 411 were delirious during their
were suffering from sustained coma in the fatigue, consisting of eight questions scored
ICU stay. Overall, 183 patients died, of
ICU; had serious auditory or visual disorders; on a 7-point Likert scale (24), was used. The
were unable to understand Dutch; were se-
whom 58 (5%) had not been delirious and
range of the CIS-fatigue is 8 to 56, a higher
verely mentally disabled; were suffering from a score indicating more pronounced fatigue.
80 (19%) had. The hypoactive delirium
serious receptive aphasia; or whose delirium The third instrument was the validated Dutch subgroup had a similar number of survi-
screening was not complete during their ICU translation (25) of the cognitive failure ques- vors compared to the mixed subgroup,
stay. Patients were diagnosed with delirium tionnaire (CFQ), which is a self-reported cog- while survival was significantly higher
when they had at least one positive CAM-ICU nitive functioning questionnaire. This ques- (p ⫽ .02) in the hyperactive subgroup
screening during their complete ICU stay, as tionnaire consists of 25 questions (14). The (Fig. 1), a median 18 months after ICU
previously described (17, 18). To secure the self-reported CFQ measures consist of four discharge. In total, 55 patients were ad-
quality of the delirium diagnosis, medical and dimensions (26) of cognition: memory, dis- mitted to the ICU more than once and 14
nursing files of all patients were also screened tractibility, social blunders, and names. Each patients were lost to follow-up.
daily for signs of delirium (19). When the files question of the CFQ was scored on a 5-point In total, there were 1,292 ICU survi-
contained signs of delirium without a positive Likert scale. The total score on the CFQ ranges vors (Fig. 1), of whom 272 patients (21%)
CAM-ICU screening, patients were additionally from 0 to 100, a higher score indicates more suffered from delirium during their ICU
screened by a delirium expert according to the self-reported cognitive dysfunction. Thus, our stay and 1,020 patients did not. In the
Diagnostic and Statistical Manual of Mental self-reported HRQoL survey consisted of a to- delirious group, seven patients (3%) with
Disorders-IV criteria (4) to rule out false neg- tal of 69 questions, which took an estimated a hyperactive subtype of delirium had one
atives and positives. In total, 17 patients 45– 60 mins to answer. positive CAM-ICU screening, and 264 pa-
(1.1%) were additionally screened this way by To guarantee the patient’s privacy, the sur- tients had at least two positive CAM-ICUs
a delirium expert. Patients with delirium were vey was sent out anonymously and numbered. during their ICU stay. Median 18 months
divided into three subtypes (3): hyperactive This allowed the primary and supervising in-
(interquartile range 15–21) after ICU dis-
delirium subtype with symptoms of hy- vestigator to match the returned survey with
charge, a total of 915 out of the 1,292
peralertness and agitation (Richmond Agita- the patient’s registry number in a separate,
tion Sedation Scale ⫹ 1/⫹4); hypoactive sub-
eligible patients (71%) returned the ques-
confidential database.
type in which the patient is hypoalert, tionnaire. Of these responders, 171 out of
lethargic (Richmond Agitation Sedation Scale 272 (63%) patients suffered from delir-
Statistical Analyses ium during their ICU stay and 744 out of
0/-3); and the alternating or mixed subtype
(Richmond Agitation Sedation Scale ⫹ 4/-3). The differences between those who suf- 1,020 (73%) did not. Seven hundred
This last subtype of delirium is characterized fered from delirium and nondelirium in ICU eighty-eight survivors completed all
by alternating symptoms of hyperactive and survivors were tested nonparametrically using questionnaires, 91% completed the SF-
hypoactive delirium. the Mann-Whitney U test. Dichotomous vari- 36, 98% completed the CIS-fatigue, and
Demographic variables as well as data of ables were tested with the chi-square test. 97% answered all the questions of the
severity of illness, delirium duration, and Since the results of our HRQoL were non- CFQ. The demographic data and illness-
Crit Care Med 2012 Vol. 40, No. 1 113Figure 1. Flowchart of included patients for the health-related quality of life (HRQoL) survey. ICU, intensive care unit.
related characteristics of the responders
and nonresponders are illustrated in Ta-
ble 1. Responders with delirium during
their ICU stay were significantly more Table 1. Demographic characteristics of responders and nonresponders
likely to be admitted for urgent reasons Responders Nonresponders
and for sepsis, were more likely to be Characteristics (n ⫽ 915) (n ⫽ 377) p
female than male, had a higher Acute
Physiology and Chronic Health Evalua- Age 65 关57–72兴 60 关47–71兴 ⬍.0001
tion II score, and their ICU and hospital Delirium (n ⫽ 272) 171 (19%) 101 (27%) .001
length of stay was significantly longer Hypoactive (n ⫽ 94) 63 (7%) 31 (8%)
Hyperactive (n ⫽ 32) 19 (2%) 13 (3%)
compared to patients that did not develop Mixed (n ⫽ 146) 89 (10%) 57 (15%)
delirium during their ICU stay (Table 2). Gender (male) 609 (67%) 231 (61%) .005
Sepsis (n) 28 (3%) 11 (3%) .53
Differences Between Delirium Urgent admission (n) 389 (43%) 204 (54%) ⬍.0001
Acute Physiology and Chronic Health 14 关11–17兴 13 关10–17兴 .06
and Nondelirium Patients on Evaluation II score
HRQoL Length of stay-intensive care unit 1 关1–2兴 1 关1–3兴 .03
(days)
SF-36. Eighteen months (median 18, Length of stay-hospital (days) 7 关5–14兴 9 关6–18兴 .001
interquartile range 15–21) after ICU dis- Admission Type ⬍.05
charge, patients with delirium during Surgical 666 (73%) 225 (59%)
their ICU stay rated their quality of life Medical 131 (14%) 74 (20%)
Trauma 41 (5%) 32 (9%)
lower on all dimensions of the SF-36 and Neurology/neurosurgical 77 (8%) 46 (12%)
the physical and mental component
scores compared to patients who did not Data are expressed as median with interquartile range unless other reported.
114 Crit Care Med 2012 Vol. 40, No. 1have delirium (Table 3). However, when differences in the CIS scores between the charge than before when compared with
adjusted for the covariates Acute Physiol- two groups remained. nondelirium patients (Table 3).
ogy and Chronic Health Evaluation II CFQ. The delirium survivors reported
score, sepsis, ICU length of stay, gender, more pronounced cognitive failure on all Duration of Delirium and
and urgent admission, no statistically sig- measured cognitive dimensions com-
HRQoL
nificant differences between groups re- pared to patients who did not suffer from
mained. The results of our ICU survivors delirium. Even after adjusting for covari- The median duration of delirium was
were worse on several domains of the ates, this difference between the groups 2 days (interquartile range 1–7, range
SF-36 compared with an age-adjusted persisted. Adjusted for covariates, pa- 1– 69 days). The delirium duration was
general Dutch population (Table 3), and tients who had previously had delirium significantly correlated with the dimen-
are in line with those of others (10). tended to experience more problems with sions memory (r ⫽ .21; p ⫽ .01) and
CIS-Fatigue. Patients who suffered their memory (p ⫽ .08). Overall, their names (r ⫽ .18; p ⫽ .04) of the CFQ. This
from delirium experienced more prob- total self-reported cognitive function was indicates that a longer duration of delir-
lems with physical exertions, expressed as significantly impaired. In addition, pa- ium is related to more pronounced prob-
a higher total CIS score, compared to the tients with delirium reported signifi- lems in memory and remembering
nondelirium patients (Table 3). Again, af- cantly more long-term problems with names. No other statistically significant
ter adjusting for covariates, no significant memory and concentration after ICU dis- correlations between duration of delir-
ium and the dimensions of the SF-36 and
CIS-fatigue were found.
Table 2. Demographic characteristics of responders
Differences in HRQoL between
Nondelirium Patients Delirium Patients
Characteristics (n ⫽ 744) (n ⫽ 171) p Subtypes of Delirium
Age 65 关57–72兴 65 关58–75兴 .13 There were no differences between the
Gender (male) 508 (68%) 101 (60%) .01 subgroups of delirium concerning age,
Acute Physiology and Chronic Health 13 关10–16兴 17 关14–20兴 ⬍.0001 Acute Physiology and Chronic Health
Evaluation II score Evaluation II score, gender, and sepsis.
Urgent admission (n) 261 (35%) 128 (75%) ⬍.0001 However, there were significant differ-
Length of stay-intensive care unit (days) 1 关1–1兴 5 关2–11兴 ⬍.0001
Length of stay-hospital (days) 7 关5–11兴 16 关9–37兴 ⬍.0001 ences between the delirium subtypes on
Sepsis (n) 12 (2%) 16 (9%) ⬍.0001 admission type, admission to the ICU for
Admission Type ⬍.01 urgent reasons, and ICU and in-hospital
Surgical 589 (79%) 77 (45%) length of stay (Table 4). These variables
Medical 77 (10%) 54 (32%)
were considered as covariates. In the un-
Trauma 24 (3%) 17 (10%)
Neurology/neurosurgical 54 (7%) 23 (14%) adjusted database, survivors of a hypoac-
tive delirium subtype evaluated their
Data are expressed as median with interquartile range unless other reported. HRQoL on several dimensions as higher
Table 3. Results of Short Form-36, checklist individual strength-fatigue, and the cognitive failure questionnaire measurements 18 months after intensive
care unit discharge adjusted for covariates
Nondelirium Patients Delirium Patients General Population Subgroup
Results (n ⫽ 744) (n ⫽ 171) pa Age 55–64 yrs (22)
Short Form-36
Physical functioning 75 关50–90兴 55 关25–80兴 .18 72 ⫾ 26
Role-physical 50 关0–100兴 25 关0–75兴 .20 67 ⫾ 41
Bodily pain 78 关57–100兴 78 关55–100兴 .26 71 ⫾ 25
General health 60 关40–75兴 55 关35–70兴 .90 62 ⫾ 20
Social functioning 88 关63–100兴 75 关50–88兴 .65 82 ⫾ 23
Vitality 60 关45–75兴 55 关40–75兴 .94 68 ⫾ 20
Role-emotional 100 关33–100兴 100 关22–100兴 .64 81 ⫾ 35
Mental health 80 关64–92兴 72 关60–88兴 .26 77 ⫾ 18
Physical component score 44 关35–52兴 38 关31–48兴 .66 50 ⫾ 9
Mental component score 53 关43–58兴 50 关38–57兴 .61 52 ⫾ 10
Checklist individual strength-total 28 关17–39兴 32 关22–44兴 .13
Cognitive Failure Questionnaire
Memory 7.0 关4–10兴 8.0 关5–12兴 .08
Distractibility 11.0 关6–15兴 11.0 关7–16兴 .19
Social blunders 6.0 关4–9兴 8.0 关4–10兴 .04b
Names 3.0 关2–4兴 3.0 关2–4兴 .22
Cognitive failure questionnaire-total 26 关17–35兴 28 关19–39兴 .03b
Data are expressed as median with interquartile range or mean with SD (⫾).
a
Adjusted for gender, urgent admission, Acute Physiology and Chronic Health Evaluation II score, sepsis, and length of intensive care unit stay using
log-transformed data (not shown); b⬍ .05.
Crit Care Med 2012 Vol. 40, No. 1 115Table 4. Differences between subtypes of delirium on health-related quality of life scores
Hypoactive Subtype Hyperactive Subtype Mixed Subtype
Health-Related Quality of Life Scores (n ⫽ 63) (n ⫽ 19) (n ⫽ 89)
Age 68 关59–75兴 64 关57–75兴 64 关57–75兴
Gender (male) 36 (57%) 10 (53%) 55 (62%)
Acute Physiology and Chronic Health Evaluation II score 16 关14–21兴 14 关13–18兴 17 关15–21兴
Urgent admission (N) 43 (68%) 9 (47%) 76 (85%)a,b
Length of stay-intensive care unit (days) 4 关2–7兴 3 关1–6兴 8 关3–16兴a,b
Length of stay-hospital (days) 15 关7–29兴 10 关5–20兴 24 关12–24兴a,b
Sepsis (n) 3 (5%) 1 (5%) 12 (14%)
Admission Type
Surgical 29 (46%) 14 (74%)c 34 (38%)a,b
Medical 21 (33%) 3 (16%)c 30 (34%)b
Trauma 4 (6%) 2 (11%)c 11 (12%)
Neurology/neurosurgical 9 (14%) 0 (0%)c 14 (16%)b
Short Form-36d
Physical Functioning 66 关35–85兴 32 关15–71兴 50 关30–75兴
Role-physical 50 关0–100兴 38 关0–100兴 25 关0–63兴
Bodily pain 78 关67–100兴 57 关32–100兴 78 关55–100兴
General health 56 关38–70兴 48 关19–65兴 50 关35–65兴
Social functioning 75 关63–100兴 63 关34–90兴 69 关50–88兴
Vitality 58 关45–76兴 50 关35–60兴 55 关40–70兴
Role-emotional 100 关33–100兴 83 关17–100兴 100 关0–100兴
Mental health 80 关65–92兴a,c 64 关56–84兴 72 关52–84兴
Physical component score 37 关22–48兴 41 关33–49兴 36 关29–45兴
Mental component score 48 关33–56兴 52 关41–59兴 49 关37–57兴
Checklist Individual Strengthd
Checklist individual strength-total 30 关16–44兴 33 关26–48兴 33 关23–44兴
Cognitive Failure Questionnaired
Memory 9 关5–12兴 8 关5–13兴 8 关5–12兴
Distractibility 11 关7–16兴 11 关6–16兴 11 关7–16兴
Social blunders 8 关4–9兴 5 关2–11兴 8 关5–11兴
Names 3 关2–4兴 4 关3–5兴 3 关2–4兴
Cognitive failure questionnaire-total 29 关20–37兴 25 关17–39兴 29 关19–42兴
a
Significant difference between hypoactive and mixed type subtype; bSignificant difference between hyperactive and mixed type subtype; cSignificant
difference between hypoactive and hyperactive subtype; dAdjusted for urgent, length of intensive care unit and in-hospital stay, and admission type using
log-transformed data (data not shown).
compared with hyperactive and mixed cognitive problems than those who did that, in addition to role functioning,
delirium survivors. After adjusting for not, even after adjusting for covariates. there was no statistically significant dif-
the covariates, patients who had a hy- Furthermore, delirium duration was sig- ference between either group (13), while
poactive delirium evaluated their men- nificantly correlated to problems with in another long-term study it was ob-
tal health significantly better than memory and names. Interestingly, after served that duration of delirium was in-
those who suffered from a mixed or adjusting for relevant covariates, survi- dependently associated with more pro-
hyperactive delirium subtype (p ⫽ .01 vors with a hyperactive or mixed subtype nounced cognitive impairment (6).
and p ⫽ .04, respectively). of delirium qualified their mental health Definite conclusions cannot be drawn from
We found no other significant differ- on the SF-36 as significantly worse than these relatively small studies because they
ences in the SF-36, CIS-fatigue, and the hypoactive delirium patients. used a more restricted HRQoL survey (13),
CFQ tests between the subtypes of de- Delirium is recognized as a frequent their maximum follow-up duration was 12
lirium. Taken together, the three sub- disorder with serious short-term health- months (6, 13), they mainly focused upon
groups of delirium suffered more exten- related problems and is associated with cognitive impairment (6), and made no ad-
sive cognitive impairment compared to longer hospital length of stay and in- justments for relevant covariates (13). This
the patients without delirium during creased mortality rates (5, 27–30). Fur- last point is of particular concern since
their ICU stay.
thermore, in long-term studies it is rec- more severely ill patients have a higher
ognized that hospitalized, non-ICU incidence of delirium and long-term im-
DISCUSSION patients with delirium suffer from persis- pairments, which may not be related to
We demonstrated that at median 18 tent cognitive impairment (31, 32). Also, each other (27).
months after ICU discharge there was no ICU patients suffer from persistent cog- The strength of the present study is
difference between delirium and nonde- nitive impairment during long-term fol- that we used a set of validated question-
lirium patients on all domains of the low-up (7, 33, 34), but in these studies, naires such as the SF-36, which is the
SF-36 and the CIS-fatigue, adjusted for no distinction between delirious and non- preferred choice for the post-ICU set-
relevant covariates. However, patients delirious patients was made. A long-term ting (20). In addition, because of the
who suffered from delirium during their ICU study that distinguished between de- large sample size we were able to cor-
ICU stay experienced significantly more lirious and nondelirious patients showed rect for covariates, and the longer fol-
116 Crit Care Med 2012 Vol. 40, No. 1low-up emphasizes the clinical rele- to our intensive care delirium protocol, In conclusion, in this large and long-
vance of the observations. patients are treated with haloperidol term follow-up study, we demonstrated
Overall and consistently, each group when a patient has at least one positive that ICU survivors with delirium during
of delirium subtype evaluated their cog- CAM-ICU screening. This early treatment their ICU stay had a similar adjusted
nitive functioning lower than the patients with haloperidol may result in negative HRQoL evaluation, but experienced sig-
who did not suffer from delirium during results in the following CAM-ICU’s. nificantly more cognitive problems in
their ICU stay. In our study, we found Therefore, to include patients with two or comparison to those who did not suffer
that patients who suffered from a hypo- more positive CAM-ICU scores may un- from delirium. Furthermore, the dura-
active delirium evaluated their HRQoL on derestimate the presence of delirium in tion of delirium was related to long-term
several domains of the SF-36 as less af- successfully treated patients (with halo- cognitive problems.
fected than the hyperactive or mixed sub- peridol). To not recognize these patients
type delirium patients. After adjusting for as delirium patients is, in our opinion, REFERENCES
relevant covariates, the domain mental not correct and not in line with daily
health remained significantly better in practice. In addition, in total, only seven 1. Ely EW, Margolin R, Francis J, et al: Evaluation
hypoactive delirium survivors. The hypo- out of the 171 responding patients with of delirium in critically ill patients: Validation
of the Confusion Assessment Method for the
active subtype is associated with a higher delirium had only one positive CAM-ICU
Intensive Care Unit (CAM-ICU). Crit Care Med
mortality rate (35, 36), a finding that we screening, and they were all treated with
2001; 29:1370 –1379
confirmed in our study, and this may haloperidol following the first positive 2. Ouimet S, Kavanagh BP, Gottfried SB, et al:
have biased the results to some extent. CAM-ICU. These were all patients with a Incidence, risk factors and consequences of
Our findings of prolonged cognitive hyperactive delirium subtype. The results ICU delirium. Intensive Care Med 2007; 33:
impairment in ICU survivors who suf- of our study would not be influenced if 66 –73
fered from delirium corroborate the re- these seven patients were not included. 3. Peterson JF, Pun BT, Dittus RS, et al: Delir-
sults of a recent meta-analysis that Thirdly, we adjusted for significant differ- ium and its motoric subtypes: A study of 614
showed that hospitalized (non-ICU) pa- ences in demographic variables between critically ill patients. J Am Geriatr Soc 2006;
tients with delirium have a significantly nondelirium and delirium patients. Since 54:479 – 484
4. American Psychiatric Association: Diagnostic
increased risk of developing dementia delirium is an independent predictor of
and Statistical Manual of Mental Disorders
(37). Our results that duration of delir- longer ICU length of stay (27), presum-
(DSM-IV). Fourth Edition. Washington DC,
ium correlates with prolonged cognitive ably independent of severity of illness, 1994
problems further extends the reported ef- then adjustment for ICU length of stay in 5. Ely EW, Gautam S, Margolin R, et al: The
fects in 77 patients 12 months after their the analyses relating delirium to long- impact of delirium in the intensive care unit
ICU stay (6) and illustrates its clinical term outcomes may underestimate the on hospital length of stay. Intensive Care
importance. This may indicate that inter- long-term effects of delirium. Further- Med 2001; 27:1892–1900
ventions aimed at reducing delirium in- more, we measured patients’ long-term 6. Girard TD, Jackson JC, Pandharipande PP, et
cidence and/or shortening its duration evaluation on HRQoL after ICU discharge al: Delirium as a predictor of long-term cog-
may produce long-term beneficial effects. once only. This can be considered as a nitive impairment in survivors of critical ill-
ness. Crit Care Med 2010; 38:1513–1520
This has not been studied yet. limitation since we do not know how
7. Hopkins RO, Jackson JC: Long-term neuro-
We wish to acknowledge several study patients’ quality of life developed during
cognitive function after critical illness. Chest
limitations. Firstly, it is intrinsic to long- these 18 months. It appears plausible that 2006; 130:869 – 878
term research in this patient group that the results would have been different if 8. Cuthbertson BH, Rattray J, Campbell MK, et
the most severely ill may not be alive 18 we would have also measured them in an al: The PRaCTICaL study of nurse led, inten-
months after their ICU discharge. Since earlier stage after discharge. Khouli et al sive care follow-up programmes for improv-
the occurrence and duration of delirium (39) showed that a higher proportion of ing long term outcomes from critical illness:
is related to increased mortality (27, 29, older patients died within 6 months after A pragmatic randomised controlled trial.
38), and the cognitive impairments re- ICU discharge, and the HRQoL worsened BMJ 2009; 339:b3723
cover in time (6), this may result in an after 6 months in the oldest group but 9. Hofhuis JG, Spronk PE, van Stel HF, et al:
The impact of severe sepsis on health-related
underestimation of the effects of delirium improved in the younger group. How-
quality of life: A long-term follow-up study.
on cognitive impairment in a long-term ever, taking into account the fact that
Anesth Analg 2008; 107:1957–1964
study such as ours. This implies that the cognitive impairment improved in delir- 10. Hofhuis JG, Spronk PE, van Stel HF, et al:
correlation between duration of delirium ium patients between 3 and 12 months The impact of critical illness on perceived
and HRQoL and cognitive impairment after ICU discharge (6), differences be- health-related quality of life during ICU
could be underestimated in our popula- tween the delirium and nondelirium ICU treatment, hospital stay, and after hospital
tion. Secondly, we diagnosed delirium survivors in our group was probably more discharge: A long-term follow-up study.
based on minimal one positive CAM-ICU pronounced earlier in the course of re- Chest 2008; 133:377–385
screening during patients’ ICU stay. One covery. Since the aim of our study was to 11. Korosec Jagodic H, Jagodic K, Podbregar M:
could argue that it is better to use at least examine the long-term effects of delir- Long-term outcome and quality of life of
patients treated in surgical intensive care: A
two consecutive positive CAM-ICU ium, we decided not to conduct repeated
comparison between sepsis and trauma. Crit
screenings to diagnose delirium. How- measures of the HRQoL status in a
Care 2006; 10:R134
ever, in all guidelines and delirium pro- smaller group of patients, instead we 12. Myhren H, Ekeberg Ø, Stokland O: Health-
tocols we are aware of, patients are chose to measure one point in time, related quality of life and return to work after
treated when they meet the criteria of after 18 months, in a large group of critical illness in general intensive care unit
delirium. This is the case following one patients. This allowed adjustment for patients: A 1-year follow-up study. Crit Care
positive CAM-ICU screening. According relevant covariates. Med 2010; 38:1554 –1561
Crit Care Med 2012 Vol. 40, No. 1 11713. Van Rompaey B, Schuurmans MJ, Short- populations. J Clin Epidemiol 1998; 51: 31. Kat MG, Vreeswijk R, de Jonghe JF, et al:
ridge-Baggett LM, et al: Long term outcome 1055–1068 Long-term cognitive outcome of delirium in
after delirium in the intensive care unit. 22. Ware JE Jr, Kosinski M, Gandek B, et al: elderly hip surgery patients. A prospective
J Clin Nurs 2009; 18:3349 –3357 The factor structure of the SF-36 Health matched controlled study over two and a half
14. Broadbent DE, Cooper PF, FitzGerald P, et Survey in 10 countries: Results from the years. Dement Geriatr Cogn Disord 2008;
al: The Cognitive Failures Questionnaire IQOLA Project. International Quality of 26:1– 8
(CFQ) and its correlates. Br J Clin Psychol Life Assessment. J Clin Epidemiol 1998; 32. MacLullich AM, Beaglehole A, Hall RJ, et
1982; 21:1–16 51:1159 –1165 al: Delirium and long-term cognitive im-
15. Ely EW, Inouye SK, Bernard GR, et al: De- 23. Ware JE, Kosinkski M, Gandek B: SF-36 pairment. Int Rev Psychiatry 2009; 21:
lirium in mechanically ventilated patients: Health Survey; Manual & Interpretation 30 – 42
Validity and reliability of the confusion as- Guide. Lincoln, RI, Quality Metric, 2005 33. Jackson JC, Hart RP, Gordon SM, et al: Six-
sessment method for the intensive care unit 24. Vermeulen RC: Translation and validation of month neuropsychological outcome of med-
(CAM-ICU). JAMA 2001; 286:2703–2710 the Dutch language version of the CDC ical intensive care unit patients. Crit Care
16. van den Boogaard M, Pickkers P, van der Symptom Inventory for assessment of Med 2003; 31:1226 –1234
Hoeven H, et al: Implementation of a de- Chronic Fatigue Syndrome (CFS). Popul 34. Jackson JC, Mitchell N, Hopkins RO: Cogni-
lirium assessment tool in the ICU can in- Health Metr 2006; 4:12 tive functioning, mental health, and quality
of life in ICU survivors: An overview. Crit
fluence haloperidol use. Crit Care 2009; 25. Merckelbach H, Muris P, Nijman H, et al:
Care Clin 2009; 25:615– 628, x
13:R131 Self-reported cognitive failures and neurotic
35. Kiely DK, Jones RN, Bergmann MA, et al:
17. Shehabi Y, Riker RR, Bokesch PM, et al: symptomatology. Personal Individual Differ-
Association between psychomotor activity
Delirium duration and mortality in lightly ences 1996; 20:715–724
delirium subtypes and mortality among
sedated, mechanically ventilated intensive 26. Wallace JC, Kass SJ, Stanny CJ: The cogni-
newly admitted post-acute facility patients.
care patients. Crit Care Med 2010; 38: tive failures questionnaire revisited: Dimen-
J Gerontol A Biol Sci Med Sci 2007; 62:
2311–2318 sions and correlates. J Gen Psychol 2002;
174 –179
18. Thomason JW, Shintani A, Peterson JF, et al: 129:238 –256
36. Yang FM, Marcantonio ER, Inouye SK, et al:
Intensive care unit delirium is an indepen- 27. Ely EW, Shintani A, Truman B, et al: Delir-
Phenomenological subtypes of delirium in
dent predictor of longer hospital stay: A pro- ium as a predictor of mortality in mechani- older persons: Patterns, prevalence, and
spective analysis of 261 non-ventilated pa- cally ventilated patients in the intensive care prognosis. Psychosomatics 2009; 50:
tients. Crit Care 2005; 9:R375–R381 unit. JAMA 2004; 291:1753–1762 248 –254
19. Inouye SK, Leo-Summers L, Zhang Y, et al: 28. Inouye SK, Rushing JT, Foreman MD, 37. Witlox J, Eurelings LS, de Jonghe JF, et al:
A chart-based method for identification of et al: Does delirium contribute to poor Delirium in elderly patients and the risk of
delirium: Validation compared with inter- hospital outcomes? A three-site epidemio- postdischarge mortality, institutionalization,
viewer ratings using the confusion assess- logic study. J Gen Intern Med 1998; 13: and dementia: A meta-analysis. JAMA 2010;
ment method. J Am Geriatr Soc 2005; 53: 234 –242 304:443– 451
312–318 29. Pisani MA, Kong SY, Kasl SV, et al: Days of 38. Lin SM, Liu CY, Wang CH, et al: The impact
20. Angus DC, Carlet J, 2002 Brussels Roundta- delirium are associated with 1-year mortality of delirium on the survival of mechanically
ble Participants: Surviving intensive care: A in an older intensive care unit population. ventilated patients. Crit Care Med 2004; 32:
report from the 2002 Brussels Roundtable. Am J Respir Crit Care Med 2009; 180: 2254 –2259
Intensive Care Med 2003; 29:368 –377 1092–1097 39. Khouli H, Astua A, Dombrowski W, et al:
21. Aaronson NK, Muller M, Cohen PD, et al: 30. van den Boogaard M, Peters SA, van der Ho- Changes in health-related quality of life and
Translation, validation, and norming of the even JG, et al: The impact of delirium on the factors predicting long-term out-
Dutch language version of the SF-36 Health prediction of in-hospital mortality in intensive comes in older adults admitted to intensive
Survey in community and chronic disease care patients. Crit Care 2010; 14:R146 care units. Crit Care Med 2011; 39:731–737
118 Crit Care Med 2012 Vol. 40, No. 1You can also read
