Intensive Care Med (2006) 32:923–926

DOI 10.1007/s00134-006-0112-y BRIEF REPORT

Christina Jones
Richard D. Griffiths
Significant cognitive dysfunction in
Tracy Slater non-delirious patients identified during and
Kirsten S. Benjamin
Sally Wilson persisting following critical illness

Received: 13 February 2006
Accepted: 13 February 2006
Published online: 8 March 2006
© Springer-Verlag 2006
C. Jones (u) · R. D. Griffiths · T. Slater
Whiston Hospital, Intensive Care Unit,
L35 5DR Prescot, UK
e-mail: christinajonesc@aol.com
Tel.: +44-151-4261600 ext. 2382
Fax: +44-151-4301628
K. S. Benjamin
Health and Safety Laboratory,
Department of Work Psychology,
Buxton, UK
S. Wilson
University of Sheffield,
Sheffield Institute for Studies on Aging,
Sheffield, UK
Abstract Objective: Recent studies
have shown significant cognitive
problems some months after critical
illness. However there has been no
research examining cognitive function
within the intensive care unit (ICU)
in non-delirious patients. Design
and setting: A prospective study in
an ICU. Patients and participants:
Using the Cambridge Neuropsy-
chological Test Automated Battery
(CANTAB), 30 long-stay, tracheal-
intubated ICU patients were tested.
Prior to testing on ICU the Confusion
Assessment Measure (CAM-ICU)
was administered and only those
patients clearly not delirious and off
sedation for several days were tested.
The CANTAB tests were repeated
a week after ICU discharge on the
general ward and then again at 2
months. Sixteen patients completed
the follow-up. Results: While on
ICU all 30 patients showed significant
problems with strategic thinking and
problem solving; 20 patients had some
problems with memory. The degree
of difficulty with problem solving on
ICU was correlated with length of
ICU stay (p = 0.011), age (p = 0.036)
and length of hospital stay post ICU
(p = 0.044). Problems with memory
in ICU and on the general ward were
correlated with admission APACHE
II score (p = 0.004 and p = 0.005 re-
spectively). At the 2-month follow-up
5 of 16 patients (31%) scored below
the 25 percentile for memory and
8 of 16 (50%) below the 25 percentile
for problem solving (Slater TA,
Jones C, Griffiths RD, Wilson S,
Benjamin K (2004) Cognitive impair-
ment during and after intensive
care: a pilot study. Intensive Care
Med 30 [Suppl 1]:S199). Con-
clusions: Difficulties with prob-
lem solving and poor memory
remained a significant issue for
2 months after ICU discharge.
Keywords ICU · Critical illness ·
Memory · Problem solving

Introduction following mechanical ventilation, compared with only 3%

Despite variation in the presenting problem of critically
ill patients, there are systemic consequences that could
compromise cognitive function. Significant numbers of
patients become septic with systemic inflammation, and
in very severe sepsis even brain cell death or oedema
may occur [1, 2]. Recent research has indicated that there
may be significant cognitive dysfunction following critical
illness [3, 4]. Cognitive dysfunction was shown in 43% of
chronic obstructive pulmonary disease (COPD) patients
in the non-ventilated controls [5]. Similarly, significant
cognitive impairment was reported in patients recovering
from adult respiratory distress syndrome (ARDS) 1 year
after ICU admission [3, 6]. Seventy-eight per cent of these
ARDS patients were shown to have some impairment of
memory, attention, concentration or mental processing
speed. Medical ICU patients present a similar picture,
with 32% of the patients having cognitive impairment at 6
months post ICU [4]. However, studies so far have been
cross-sectional and have not looked at the evolving impact
924

on the patient during the ICU stay or on the general ward excluded from our study because they were delirious
after ICU discharge. when tested with the CAM-ICU. Thirty long-stay ICU
patients, shown to be not delirious were tested using the
CANTAB. Sixteen of these 30 patients completed the
2-month follow-up. Two patients died after discharge, and
Material and methods
there was a tendency for the younger patients to drop out
Long-stay ICU patients (stay greater than 6 days, venti- (median age 46 years for the drop-outs versus 54 years
lated for more than 48 h), off all sedative agents and opi- for those completing the 2-month follow-up). In addition,
ates for at least 3 days, were tested with the Confusion patients with significant problems with memory on the
Assessment Measure (CAM-ICU) [7]. The patients were pattern recognition test while in ICU were more likely to
approached to enter the study only if they were found not drop out of the study (p = 0.046). No patients recruited
to be delirious and no fluctuation in mental alertness had to the study had a previous history of neurological or
been noted. The requirement to have been off all sedative psychological disorders.
or analgesic drugs for 3 days but still be in ICU resulted
in recruitment of only patients with a tracheostomy, rather
than oral intubation. A history was taken from the medical
notes detailing any psychological and medical problems Within ICU
at admission to hospital. The patients were then tested on
ICU using the Cambridge Neuropsychological Test Auto- While on the ICU all 30 patients showed significant prob-
mated Battery (CANTAB). This is a computerised system lems with strategic thinking and problem solving in terms
using a touch screen and is completely non-verbal. It was of being able to follow a set of rules to solve a visual puz-
chosen over other neuropsychological test batteries as all zle, scoring ≤ 25 percentile compared with an age-, sex-
the patients tested had a tracheostomy and could not talk matched control population. Eight patients (27%) could
to answer questions; no other battery of tests allowed this. not solve any of the problems. Twenty patients (66.7%)
The CANTAB is also able to take into account physical had some problems (scoring ≤ 25 percentile) with mem-
weakness and subtract physical reaction time from the to- ory using the pattern recognition test. Yet these patients
tal reaction time to give a true measure of thinking time. were neither delirious nor had they any neurological signs
The CANTAB has been used with a number of different giving rise to any clinical concern.
patient groups [8]. Consent was obtained by the patient Impairment of problem solving on ICU (n = 30)
nodding their head to verbal explanation of the study as was correlated with length of ICU stay (p = 0.011),
most could not write. A subsequent written informed con- age (p = 0.036) and length of hospital stay post ICU
sent was obtained after ICU discharge. Approval for the (p = 0.044). However, impairment of memory was corre-
study was obtained from the local ethics committee. lated with admission APACHE II score in ICU (p = 0.004)
On the basis of previous studies it was decided that two and not related to length of ICU stay or age.
important areas to test were problem solving and memory.
The test for strategic thinking and problem solving on the
CANTAB is the “Stocking of Cambridge”. This is an in-
creasingly complicated visual pattern that the patient must One week post ICU discharge
reproduce by moving coloured balls on the screen. The test
for memory on the CANTAB is pattern recognition recall. On the general ward 1 week post ICU discharge, 26
Patients were instructed to try to remember a series of pat- patients still showed significant problems with strategic
terns and then say which pattern they had seen before out thinking and problem solving. Six patients (20%) still
of two presented to them. This battery of tests was delib- could not solve any of the problems. Fifteen patients
erately kept short so that fatigue and concentration prob- (50%) still had some problems with memory. Impairment
lems would be minimised while patients were on ICU. The of memory was correlated with admission APACHE II
CANTAB was repeated on the general ward at 1 week post score (p = 0.005).
ICU discharge and in the outpatient clinic at 2 months post
ICU discharge.
Two-month follow-up
Results
At the 2-month follow-up 5 of 16 patients (31%) scored
The study period was from March 2003 to November below the 25 percentile for memory and 8 of 16 (50%) for
2004. During that period, 151 patients who stayed on problem solving (see Fig. 1). Impairment of problem solv-
ICU for more than 6 days and were ventilated for more ing and memory remained a significant issue for 2 months
than 48 h were admitted. The majority of these were after ICU discharge.
 925

Table 1 Patient descriptors (n = 30)

Admission variables median (range)

Males/females 17:13
ICU stay (days) 14 (6–45)
Age (years) 54 (18–78)
APACHE II score 16 (6–25)
Diagnostic groups (no. of patients)
Peritonitis 10
Pneumonia/empyema 7
Asthma/COPD 4
Sepsis 5
Other 2
ARDS 1
Trauma 1

likely not to turn up for their follow-up test would fit

Fig. 1 Stocking of Cambridge (SOC) test of problem solving in pa-
tients with complete 2-month follow-up (n = 16)
Discussion
We have shown that patients who are not delirious still
have significant problems with problem solving and
memory while on the ICU and even on the general ward
afterwards. The observation that impairment of problem
solving was correlated with length of ICU and hospital
stay is similar to the findings of a study on delirium [9].
It has been suggested that some critically ill patients
may develop a dysexecutive syndrome [10], which may
vary in severity and result in problems with memory,
problem solving and making social decisions. Patients
themselves or their family may be aware of an increase
in “slips of action” and absentmindedness. Slips of action
are produced when we forget some part of the plan for
a sequence of actions. The observation that patients who Acknowledgements. We would like to acknowledge the help of
had significant memory problems in ICU were those more Caroline Yaro with data collection.
with this. Ability to remember appointments is likely to
be affected. Some patients were tested at home, where
practical, as they kept forgetting to attend appointments.
In the model of working memory proposed by Baddeley
and Hitch [11] the central executive forms the interface
between long-term memory and several slave systems and
is responsible for the selection and operation of strategies
and maintaining and switching attention as needed. The
central executive is assumed to coordinate information
from a range of sources. It is associated with the frontal
lobes and is sensitive to frontal lobe damage, producing
the dysexecutive syndrome [10]. Other studies have shown
a similar deficit in executive function [12].
The mechanisms behind the cognitive deficits demon-
strated by patients recovering from critical illness are not
understood at present and may be multifactorial. There is
a need to take these previously unrecognised problems in
our patients seriously. This new work raises the possibility
that these patients may require more detailed investigations
to see whether a treatable condition exists.

References
1. Sharshar T, Gray F, Lorin de la Grand-
maison G, Hopkinson NS, Ross E,
Dorandeu A, Orlikowski D, Raphael
J-C, Gajdos P, Annane D (2003)
Apoptosis of neurons in cardiovas-
cular autonomic centres triggered
by inducible nitric oxide synthase
after death from septic shock. Lancet
362:1799–1805
2. Zauner C, Gendo A, Kramer L, Funk
GC, Bauer E, Schenk P, Ratheiser K,
Madl C (2002) Impaired subcortical
and cortical sensory evoked potential
pathways in septic patients. Crit Care
Med 30(5):1136–1139
3. Hopkins RO, Weaver LK, Pope D,
Orme JF, Bigler ED, Larson-Lohr V
(1999) Neuropsychological sequelae
and impaired health status in survivors
of severe acute respiratory distress
syndrome. Am J Respir Crit Care Med
160:50–56
926
4. Jackson JC, Hart RP, Gordon SM,
Shintani A, Truman B, May L, Ely
EW (2003) Six month neuropsycholog-
ical outcome from medical intensive
care unit patients. Crit Care Med
31(4):1226–1234
5. Ambrosino N, Bruletti G, Scala V, Porta
R, Vitacca M (2000) Cognitive and
perceived health status in patients with
chronic obstructive pulmonary disease
surviving acute on chronic respiratory
failure: a controlled study Intensive
Care Med 28:170–177
6. Hopkins RO, Weaver LK, Chan KJ, 8. Downes JJ et al (1989) Impaired
Orme JF (2004) Quality of life, emo- extra-dimensional shift performance in
tional and cognitive function following medicated and unmedicated Parkinson’s
acute respiratory distress syndrome. J disease: evidence for a specific atten-
Int Neuropsychol Soc 10(7):1005–1017 tional dysfunction. Neuropsychologia
7. Ely W, Margolin R, Francis J et al 27:1329–1343
(2001) Evaluation of delirium in crit-
ically ill patients: validation of the
Confusion Assessment Method for the
Intensive Care Unit (CAM-ICU). Crit
Care Med 29:1370–1379