Article Analysis 1

B

W m

L M A Å K J a

I b

c

d

e

f

A a

a

A R R 3 A

K M D P C E

h 1

Australian Critical Care 29 (2016) 5–14

Contents lists available at ScienceDirect

Australian Critical Care

j o u r n a l h o m e p a g e : w w w . e l s e v i e r . c o m / l o c a t e / a u c c

est nursing review paper

hat is the relationship between elements of ICU treatment and emories after discharge in adult ICU survivors?

eanne M. Aitken PhD, RN a,b,c,∗, aria I. Castillo PhD, RN a,b,

manda Ullman MAppSci, GCPICU, RN a, sa Engström PhD, RN, CCN d, athryn Cunningham PhD, MSc, MA (Honours) e,

anice Rattray PhD, MN, RGN, SCM f

School of Nursing & Midwifery & NHMRC Centre of Research Excellence in Nursing (NCREN), Centre for Health Practice Innovation, Menzies Health nstitute Queensland, Griffith University, Australia Intensive Care Unit, Princess Alexandra Hospital, Australia School of Health Sciences, City University London, UK Division of Nursing, Department of Health Science, Luleå University of Technology, Sweden Population Health Sciences, Medical Research Institute, University of Dundee, UK School of Nursing and Midwifery, University of Dundee, UK

t the conclusion of this article a Continuing Professional Development activity is ttached

r t i c l e i n f o r m a t i o n

rticle history: eceived 6 July 2015 eceived in revised form 0 November 2015 ccepted 30 November 2015

eywords: emory elusion sychological recovery ritical care vidence based nursing

a b s t r a c t

Objectives: Patients admitted to an intensive care unit (ICU) often experience distressing memories during recovery that have been associated with poor psychological and cognitive outcomes. The aim of this literature review was to synthesise the literature reporting on relationships between elements of ICU treatment and memories after discharge in adult ICU survivors. Review method used: Integrative review methods were used to systematically search, select, extract, appraise and summarise current knowledge from the available research and identify gaps in the literature. Data sources: The following electronic databases were systematically searched: PubMed, Ovid EMBASE, EBSCOhost CINAHL, PsycINFO and Cochrane Central Register of Controlled Trials. Additional studies were identified through searches of bibliographies. Original quantitative research articles written in English that were published in peer-review journals were included. Review methods: Data extracted from studies included authors, study aims, population, sample size and characteristics, methods, ICU treatments, ICU memory definitions, data collection strategies and findings. Study quality assessment was based on elements of the Critical Appraisal Skills Programme using the checklists developed for randomised controlled trials and cohort studies. Results: Fourteen articles containing data from 13 studies met the inclusion criteria and were included in the final analysis. The relatively limited evidence about the association between elements of ICU treatment and memories after ICU discharge suggest that deep sedation, corticoids and administration of glucose 50% due to hypoglycaemia contribute to the development of delusional memories and amnesia

of ICU stay. Conclusions: The body of literature on the relationship between elements of ICU treatment and memories after ICU discharge is small and at its early stages. Larger studies using rigorous study design are needed in order to evaluate the effects of the ICU during recovery.

© 2015 Australian College

∗ Corresponding author. Tel.: +61 7 3176 7257. E-mail address: [email protected] (L.M. Aitken).

ttp://dx.doi.org/10.1016/j.aucc.2015.11.004 036-7314/© 2015 Australian College of Critical Care Nurses Ltd. Published by Elsevier Lt

of different elements of ICU treatment on the development of memories

of Critical Care Nurses Ltd. Published by Elsevier Ltd. All rights reserved.

d. All rights reserved.

6 an Cri

1

e f s a e m i c r t s p o

o c t b t o t s I t t f h p m f i s s

a m e e f t a t a e m t t I

2

s

2

m m i i c

L.M. Aitken et al. / Australi

. Introduction

Patients admitted to an intensive care unit (ICU) often experi- nce distressing memories during recovery. Specifically, delusional, actual and emotional memories are frequently reported.1–3 Delu- ional memories correspond to the recall of unreal events such s hallucinations, nightmares and paranoia, which have been stimated to be present in about 20–48% of patients.3,4 Factual emories are the recall of real events that occurred during patient’s

ntensive care treatment such as the presence of an endotra- heal tube and mechanical ventilation; proportions of patients who ecall factual memories vary significantly from 18% to 96%.1,2 Emo- ional memories involve the recall of feelings such as anxiety, fear, uffocation and pain and have been reported by highly variable pro- ortions of patients, ranging from 9% to 88%.1,2,5,6 Lack of memory f ICU events has also been reported in about 18–38% of patients.1,7

Memories of ICU treatment play a significant role in the devel- pment of post-intensive care syndrome (PICS), a syndrome that is haracterised by “new or worsening impairments in physical, cogni- ive or mental health status arising after a critical illness and persisting eyond acute care hospitalisation” (page 4).8 Memories are thought o specifically affect the psychological and cognitive components f recovery of ICU survivors.9,10 For instance, delusional and emo- ional memories have been associated with the development of ymptoms of anxiety, depression and posttraumatic stress after CU discharge.11–17 The role of factual memories is unclear, with hem being identified as protecting patients from anxiety and post- raumatic stress symptoms in some cohorts,11 while in others actual memories have been associated with poorer psychological ealth during recovery.18 The number of distressing memories that atients recall was identified as a significant factor for posttrau- atic stress symptoms.6,16,19,20 The association between cognitive

unctioning and memories of the ICU has also been explored. An mproved cognitive functioning after ICU discharge was found to be ignificantly associated with having no recollections of the inten- ive care experience.21

A range of elements of ICU treatment have been proposed as ffecting psychological health, including the number and type of emories of ICU. These elements of care have included specific cat-

gories of medications such as anti-inflammatory medications, for xample hydrocortisone22–24 and sedation and analgesic agents, or example midazolam and opioids.1,2,25 Further, a link between he level of sedation and psychological health has been proposed, lthough the evidence of that relationship remains unclear.26 Given he potential influence of aspects of ICU treatment on memories, nd the link between memories of ICU and PICS, it is appropriate to xplore these links with a view to adapting our practice to improve emories. A review addressing this topic could not be located in

he current literature. The aim of this literature review was to syn- hesise the literature examining relationships between elements of CU treatment and memories after discharge in adult ICU survivors.

. Method

Integrative review methods were used to systematically search, elect, extract, appraise and synthesise the available research.27

.1. Eligibility criteria

Primary research articles were included in the review if they easured the relationship between specific ICU treatments and

emories reported by adult ICU survivors. Studies were excluded

f they were not written in English. ICU treatments were defined as nterventions administered to patients during admission to a criti- al care unit, e.g. mechanical ventilation, use of invasive devices and

tical Care 29 (2016) 5–14

administration of medications. ICU memories were defined as per the study authors, and included factual and delusional memories of the survivors’ time in a critical care unit.11

2.2. Search methods

PubMed, Ovid EMBASE, EBSCOhost CINAHL, PsycINFO and Cochrane Central Register of Controlled Trials were systematically and independently searched in May 2015. Medical Subject Head- ings (MeSH) were amnesia, memory, intensive care units, critical care, critically ill, critical illness, and intensive care. Additional stud- ies were identified through searches of bibliographies. Searches were performed without year restrictions but were limited to human studies. Titles and abstracts were scanned for relevance and eligibility using the a priori eligibility criteria. The search was undertaken by one author (AU) using search terms developed by the review team. Selection of articles based on the inclusion and exclusion criteria was completed by two authors (AU and LMA) independently, with results compared and disagreements discussed and resolved by the whole team.

2.3. Data extraction and quality appraisal

A data extraction form was developed by the study authors (AU and LMA) and applied to each of the included studies. For each paper the author, study objective, population, sample size and char- acteristics, methods, ICU treatments, ICU memory definitions, data collection strategies, findings and study quality were extracted by study authors (MIC, AU, KC). Study quality assessment, including the elements of validity, significance and usefulness, was based upon elements of the Critical Appraisal Skills Programme (CASP) checklists relevant to each included study.28

2.4. Data synthesis

Data from the included studies were categorised and summari- sed to product a coherent and logical summary across the different categories of ICU treatment.

3. Results

Following database and bibliographic searching 2748 titles were identified. This number was reduced to 1548 titles after dupli- cates were removed (Fig. 1).29 The abstracts of these titles were reviewed and 64 full text articles examined. A further 50 articles were excluded because they did not focus on the review question, with 14 articles (13 studies; one duplicate publication with some additional results2,18) included in the synthesis of results. No meta- analyses were able to be undertaken due to the diversity of designs, interventions examined and instruments used to measure mem- ories; this resulted in the absence of multiple studies examining sufficiently similar questions to enable data to be combined.

Randomised controlled trials and cohort studies were the most commonly used designs in the 13 included studies (Table 1). Sam- ple sizes were highly variable, ranging from 11 to 313 participants. Instruments used also varied widely and follow-up was conducted between 3 days and 5 years after ICU discharge. Studies gener- ally had variable levels of both bias and usefulness, with this latter aspect often limited by small sample sizes (Table 2).

Sedation, both in regard to the agents used and the depth

of sedation, was the primary intensive care intervention exam- ined in relation to memories after ICU and was the focus of eight studies.1–3,17,18,30–33 Consequently, the most common theme that was identified focused on sedation and analgesia. The effect of other

L.M. Aitken et al. / Australian Critical Care 29 (2016) 5–14 7

Table 1 Relationships between ICU treatment and memories of ICU survivors.

Author details Aim Participants and design Outcome measure Results Comments including strengths and limitations

Henderson et al., 1994; Australia33

To assess the effects of adding low dose midazolam to papaveretum on memory recall and duration of ventilation in drug paralysed post-operative patients

100 post-operative patients; blinded RCT; recall assessed just prior to ICU discharge

Locally developed ‘experience’ questionnaire (no validation)

No difference in recall in regard to pain, noise, anxiety, discomfort, memory. Duration of ventilation no different (25 vs. 26 h, p > 0.05).

No control of other aspects of care; asked patients to recall ‘period of drug paralysis’ so their recall might relate to any period of ICU care; conducted in a time where routine paralysis was common and therefore not relevant to the current practice environment.

Capuzzo et al. 2001; Italy1

To investigate the relationship between analgesia, sedation and memory of intensive care

152 ICU patients with LOS > 24 h; prospective cohort study with follow-up 6 months after hospital discharge in 1 hospital

Structured interview to assess memories (factual, sensation and emotional) based on the ICU-MT. Quality of life (locally developed instrument)

No significant difference in the incidence of factual, sensation and emotional memories between the 3 sedation groups (Group A: no morphine/0–2 doses benzodiazepines; Group B: morphine without other sedatives; Group C: morphine and other sedatives). Although bivariate analysis indicated patients reporting at least 1 emotional memory were more likely to be female, emergency admission, have infection/sepsis and receive corticosteroids, only gender was significant in logistic regression.

48 eligible patients not interviewed due to loss-to-follow-up (36), terminally ill (4) and cognitive impairment (8)–non-participants were more likely to be post-surgery with a longer ICU and hospital LOS; convenience sampling used for recruitment in single centre.

Kress et al. 2003, USA31

To search for evidence that daily interruption of sedation was associated with long-term psychological harm.

32 mechanically ventilated medical ICU patients; participants recruited from previous RCT as well as contemporaneous cohort were followed up 6 months after discharge

Structured interview by clinical psychologist plus self-report measures (IES-R, SF-36, STAI, BDI, PAIS) Locally developed questions to assess recollection of ICU

Many patients recalled being in ICU when questioned in their hospital stay (68% control vs. 69% intervention, p = 1.0); there was a trend towards more patients in the control group recalling waking in ICU when questioned at 6 months (26% control vs. 0% intervention, p = 0.06).

Specific methodology or instruments were not used to measure memories of ICU – the information appears to have been collected during the follow-up interview; potentially biased cohort given small proportion of eligible patients enrolled.

Pierce et al. 2004; United Kingdom4

To examine the association of delusional and real memories with pre-operative and post-operative factors.

161 cardiac surgical patients with ICU LOS >4 days; retrospective cohort study with follow-up 2–5 years after surgery in 1 hospital

Modified ICU-MT with no validation of modification

Patients were categories as ‘dreamers’ (1 or more memories of dreams, nightmares, thoughts that others were trying to inflict harm, were plotting against patient or that patient had travelled after surgery) or ‘non-dreamers. Factors positively associated with ‘dreamers’ included treatment with intravenous 50% glucose, midazolam, steroid therapy and episodes of sepsis, with the development of new neurological signs exerting a protective effect.

Clinical factors collected through retrospective chart review; 161 of 423 possible patients recruited (89 died, 59 whose GP refused assent, 90 no response); variable follow-up time frame.

Schelling et al. 2004; Germany23

To examine whether stress doses of hydrocortisone after cardiac surgery reduce long term incidence of chronic stress, PTSD and traumatic memories

91 cardiac surgery patients; RCT with follow-up at 2–3 days, 1 week and 6 months (n = 48) after ICU in 1 hospital

PTSS-10 No significant difference in number and categories of traumatic memories between patients in hydrocortisone and control groups: nightmares (23% vs. 36%, p = 0.36), pain (19% vs. 9%, p = 0.43), respiratory distress (19% vs. 27%, p = 0.73), anxiety/panic (31% vs. 40%, p = .33).

Participants and clinical staff blinded to group allocation; validated traumatic memories questionnaire.

8 L.M. Aitken et al. / Australian Critical Care 29 (2016) 5–14

Table 1 (Continued)

Author details Aim Participants and design Outcome measure Results Comments including strengths and limitations

Ringdal et al. 2006; Sweden34

To describe trauma patients’ memories of ICU and identify factors associated with delusional memories

239 trauma ICU patients; prospective cohort study with follow-up 6–18 months after ICU discharge in 5 hospitals

ICU-MT Patients with clear recollection of ICU had shorter ICU LOS, were less likely to require mechanical ventilation and have shorter duration of mechanical ventilation, and were less likely to receive sedatives. Patients with delusional memories were younger, had longer ICU LOS, higher temperature; lower haemoglobin and more likely to have had renal failure, surgery, ventilator support, sedatives and analgesics.

239 of 344 eligible participants recruited (66 did not reply, 39 declined; non-responders had shorter ICU LOS but otherwise did not differ from the final participants); analysis was limited to bivariate analysis with no multivariable analysis reported.

Weis et al. 2006; Germany35

To determine whether hydrocortisone administration reduced chronic stress symptoms after cardiac surgery

36 high risk cardiac surgery patients; RCT with 6 month follow-up after ICU in 1 hospital

PTSS-10 SF-36

Patients who received hydrocortisone had a shorter ICU LOS, lower TISS scores, required less norepinephrine and a trend towards lower pro-inflammatory cytokine IL-6 as well as higher quality of life scores and lower chronic stress symptom scores. Number and type of traumatic memories did not differ between the two groups (p ≤ .33).

Participants and clinical staff were blinded to group allocation; 28/36 patients followed up (2 incomplete data, 6 did not return questionnaire) with no differences between those who completed the study or not; previously validated memory instrument used.

Samuelson et al. 2006 and Samuelson et al. 2007; Sweden2,18

To investigate the relationship between (i) memory and intensive care sedation (ii) recall of stressful experiences and intensive care sedation

313 intubated mechanically ventilated adults admitted to ICU for >24 h; prospective cohort study with follow-up 6–10 days after ICU in 2 hospitals

ICU-MT ICU-SEQ

Patients with no recall of ICU (18%) were older and had fewer periods of wakefulness (MAAS 0–2) than those with memories of ICU (82%). Patients with delusional memories (34%) had longer ICU stay, higher baseline severity of illness, higher proportions of MAAS 4–6 and more midazolam than those with recall of ICU but no delusional memories. Patients with more periods of wakefulness (MAAS 3), longer ICU stay and emergency admissions recalled stressful experiences as more bothersome.

250 of 313 patients completed the study; convenience sampling used to recruit participants; patients who were lost to follow-up were more frequently emergency admissions and older than those retained in the study; previously validated ICU Memory Tool used to measure recall; follow-up only 6–10 days after ICU.

Samuelson et al. 2008; Sweden3

To assess the presence of stressful memories in light vs. heavy sedation

36 mechanically ventilated post-operative patients; RCT with 2 month follow-up after ICU in 1 hospital

ICU-MT ICU-SEQ IES-R

No significant difference in memory between light (MAAS 3–4) and heavy (MAAS 1–2) sedation patients; trend towards more delusional memories in the heavy sedation group (33% vs. 6%, p = 0.09); analysis excluding prolonged ICU stay showed higher prevalence of delusional memories in heavy sedation group (31% vs. 0%, p = 0.04).

Previously validated ICU Memory Tool used to measure recall; follow-up only 5 days after ICU; pilot study with small participant numbers.

L.M. Aitken et al. / Australian Critical Care 29 (2016) 5–14 9

Table 1 (Continued)

Author details Aim Participants and design Outcome measure Results Comments including strengths and limitations

Weinert et al. 2008; USA7

To determine the relationship between critical illness factors and ICU recall and symptoms of post-traumatic stress disorder

277 adult ICU patients; prospective cohort study with follow-up 2 and 6 months post ICU discharge

ICU amnesia score (developed by study authors–limited validation of this tool); Posttraumatic stress diagnostic scale

Intensity of sedation administration was not associated with ICU recall although there was weak association between increased wakefulness during mechanical ventilation and factual ICU recall (r2 = 0.03–0.11, p < 0.05).

Only 90 of 277 patients provided data for 2 and 6 month follow-up; those who completed 2 month follow-up were more likely to be treated in the surgical ICU, had shorter duration of mechanical ventilation and better mental status prior to intubation; recall of ICU experience measured using appears to have been locally developed with no validation described.

Sackey et al. 2008; Sweden17

To compare memories of ICU after sedation with intravenous midazolam or inhaled isoflurane

40 mechanically ventilated general ICU patients; RCT with follow-up 6 months after ICU in 1 hospital

ICU-MT HADS IES WB

Trend towards less memories of hallucinations or delusions in the isoflurane group although this did not reach statistical significance (2/10 vs. 5/7, p = 0.06). No significant differences between the groups in regard to memories of feelings or factual events.

Only 17 of 40 patients provided data (11 died; 12 non-responders); no control of other related factors such as opioid medications and ICU LOS.

Treggiari et al. 2009; Switzerland32

To determine if light sedation, compared to deep sedation, affects subsequent patient mental health

137 patients requiring mechanical ventilation (129 included in analysis); RCT with follow-up at hospital discharge and 4 weeks

PCL IES-R HADS

At hospital discharge more patients in the deep sedation group had “trouble remembering important parts of the stressful experience” (37% vs. 13%, p = 0.01), this remained similar (37% vs. 14%, p = 0.02) at 4 weeks; similar patterns were reported in regard to “repeated, disturbing memories of the stressful experience” (18% vs. 4%; p = 0.05 at both discharge and 4 weeks)

No specific instruments used to assess patients’ memories however 2 items in the PCL address memories.

Ethier et al. 2011; Canada30

To evaluate recall of ICU stay in patients managed with 2 sedation strategies: a sedation protocol or a combination of sedation protocol and daily sedation interruption

21 adult ICU patients managed with sedation protocol or no sedation protocol; pilot RCT with follow-up 72 h after ICU discharge

Patients Recall Questionnaire (develop by study authors–no validation of this tool)

No significant differences in the recall of ICU experiences between the 2 groups. More than 50% of patients in both groups recalled experiencing pain, anxiety or fear while in the ICU and 48%, 33% and 29% of the 21 patients had no memories of endotracheal tube suctioning, being on a breathing machine or being bathed, respectively.

Convenience sampling, with 26 of a potential 39 patients approached and 21 patients enrolled; Non-validated, locally developed, instrument used to measure short-term recollection of ICU; Extremely short-term (72 h) follow-up.

MAAS: Motor Activity Assessment Scale; ICU-MT: ICU Memory Tool; IES-R: Revised Impact of Event Scale; SF-36: Medical Outcomes Study SF 36 item short-form health s AIS: P s rienc P

m e

3

a o e

urvey; STAI: State and Trait Anxiety Inventory; BDI: Beck Depression Inventory; P cale; IES: Impact of Event Scale; WB: Well-Being Index; ICU-SEQ: ICU Stressful Expe CL: PTSD Checklist

edications and the duration of mechanical ventilation were also xplored in a small number of studies.

.1. Sedation and analgesia

Sedation and analgesia have been explored as possible factors ssociated with patients’ ability to recall ICU events with most f the evidence indicating that these therapies have some influ- nce in this area.2–4,17,32,34 Depth of sedation, as measured using

sychological Adjustment to Illness Score; HADS: Hospital Anxiety and Depression es Questionnaire; PTSS-10: Posttraumatic Stress Symptoms 10-Question Inventory;

various sedation assessment scales, has been proposed as hav- ing a significant impact on patients’ recall of ICU events. Deeply sedated patients reported amnesia of their ICU stay (OR 1.60, 95% CI 1.35–1.91, p < 0.0001),2 had “trouble remembering important parts of the stressful experience” (37% vs. 13%, p = 0.01),32 and reported

more repeated, disturbing memories (18% vs. 4%, p < 0.05).32 Fur- ther, more deeply sedation patients reported delusional memories (33% vs. 6%, p = 0.09)3 (OR 1.76, 95% CI 1.14–2.72, p = 0.008)2 3–5 days after ICU discharge.

10 L.M. Aitken et al. / Australian Critical Care 29 (2016) 5–14

Records id en�fied through database searc hing

(n = 2,748)

Sc re

en in

g In

cl ud

ed

El ig

ib ili

ty

Id en

� fic

a� on

Addi�onal reco rds ide n�fi ed through other sources

(n = 3)

Records a�er du plicat es rem oved (n = 1,5 48)

Reco rds screen ed (n = 1,548)

Recor ds excluded (n = 1,484)

Full-text ar�cles ass ess ed for eligibility

(n = 64)

Full-text ar�cl es exclud ed, with reason s

(n = 50) Did not measure asso cia�o n = 24

ICU tr eatment not studied = 25

Short term memo ry = 1

Ar�cles included in review (n = 14 re presen�ng 13

studies) Randomised con trol led

trials = 7 Prospec�ve cohort = 5

�ve c

A flo

p t s q a s m a (

a o c s m p r f z S m i t d m s b a

Retrospec

Fig. 1. PRISM

In contrast, the depth of sedation appeared to influence patients’ erception of stressful experiences during ICU treatment with he more awake patients reporting more stressful or bother- ome memories, as described on the ICU-stressful experiences uestionnaire.18 These findings are not consistent with Weinert nd Sprenkle7 who identified no association between inten- ity of sedation administration and ICU recall in a cohort of edical-surgical ICU patients, although they did report a weak

ssociation between increased wakefulness and factual ICU recall r2 = 0.03–0.11, p < 0.05).7

The association between specific medications such as benzodi- zepines, opioids and propofol during ICU treatment and memories f ICU has also been investigated although again, the evidence is onflicting. A retrospective study including 161 cardiac surgery urvivors reported a significant association between delusional emories and midazolam infusion (OR 3.51 95% CI 1.59–7.75,

= 0.002) 2–5 years post ICU discharge.4 On the contrary, an explo- ative multicentre investigation including 239 trauma participants ound no such relationship between delusional memories and ben- odiazepines, opioids, and propofol in multivariable analysis.34

imilarly, no statistically significant difference was reported in emories by 17 patients enrolled in a pilot RCT designed to exam-

ne the effect of midazolam vs. isoflurane on memories, although here was a trend towards less memories of hallucinations or elusions in the group treated with isoflurane.17 Further, in 152

edical-surgical ICU patients the influence different regimens of

edation and analgesia (Group A: no morphine and <2 doses of a enzodiazepine; Group B: morphine without other sedative drugs, nd Group C: morphine and other sedative drugs) was examined.1

ohort = 1

w diagram.

It was concluded that analgesia (morphine) and sedation (propofol, benzodiazepines and promazine) did not influence the incidence of factual, sensation, and emotional memories of this cohort.1

The discrepancies in these different results might be explained by the differences in the design of these studies (prospective vs. retrospective cohorts vs. pilot RCT) and the characteristics of the samples investigated (trauma, cardiac surgery and general ICU patients). For example, delusional memories were reported by 26% of the trauma patients34 compared to 48% of the cardiac surgery group.4 In addition, benzodiazepines (no specific information pro- vided about what medications were used) were administered to 24% of the trauma patients34 compared to 32% of the cardiac surgery group who received midazolam.4 Delusional memories were not specifically reported in the medical-surgical ICU group1 but emo- tional memories incorporating hallucinations, nightmares, dreams and feeling confused or down were reported in 15% of the group receiving <2 doses of benzodiazepines and 32% of the group receiv- ing unlimited sedatives. In the cardiac surgery cohort both the prevalence of delusional memories (48%) and administration of midazolam (32%) were much higher than the trauma cohort.4 Of note, the trauma patients had an average ICU LOS of approximately 4 days34 while the cardiac surgery patients remained in ICU for an average of 5 (non-dreamers) or 7 days (dreamers).4

Different sedation strategies such as sedation protocols or seda- tion interruption have been proposed as influencing patients’ recall

of their time in ICU. Despite the theoretical basis for such links, no difference in the recall of ICU experiences including recollections of fear, anxiety and pain measured on a locally developed instrument were reported by 21 patients enrolled in a pilot RCT.30 This pilot

L.M. Aitken et al. / Australian Critical Care 29 (2016) 5–14 11

Table 2 Study quality appraisal using CASP criteria.

Author details Method Validity Significance Usefulness

Henderson et al., 199433

RCT Selection bias: Unclear Measurement bias: High Assessment of confounding variables: Low Longevity of follow-up: Low

Reporting bias: Unclear Imprecise results

Low usefulness

Capuzzo et al. 20011

Prospective cohort Selection bias: Low Measurement bias: Unclear Assessment of confounding variables: High Longevity of follow-up: Unclear

Reporting bias: Low Precise results

Low usefulness

Kress et al. 200331 Prospective and retrospective cohort

Selection bias: Low Measurement bias: High Assessment of confounding variables: Low Longevity of follow-up: Low

Reporting bias: Low Imprecise results

Low usefulness

Pierce et al. 20044 Retrospective cohort

Selection bias: High Measurement bias: High Assessment of confounding variables: Unclear Longevity of follow-up: High

Reporting bias: Low Imprecise results

Low usefulness

Schelling et al. 200423

RCT Selection bias: Low Performance bias: Low Attrition bias: Low Detection bias: Low

Reporting bias: Low Imprecise results

Moderate usefulness

Ringdal et al. 200634

Prospective cohort Selection bias: Low Measurement bias: Low Assessment of confounding variables: Low Longevity of follow-up: Low

Reporting bias: Low Precise results

High usefulness

Weis et al. 200635 RCT Selection bias: Low Performance bias: Unclear Attrition bias: Low Detection bias: Low

Reporting bias: Low Imprecise results

Moderate usefulness

Samuelson et al. 2006 and Samuelson et al. 20072,18

Prospective cohort Selection bias: Unclear Measurement bias: Low Assessment of confounding variables: Low Longevity of follow-up: High

Reporting bias: Low Precise results

Moderate usefulness

Samuelson et al. 20083

RCT Selection bias: Low Performance bias: Unclear Attrition bias: low Detection bias: Low

Reporting bias: Low Imprecise results

Moderate usefulness

Weinert et al. 20087

Prospective cohort Selection bias: Unclear Measurement bias: Unclear Assessment of confounding variables: Low Longevity of follow-up: Low

Reporting bias: Low Precise results

Low usefulness

Sackey et al. 200817 RCT Selection bias: Unclear Performance bias: High Attrition bias: Low Detection bias: Low

Reporting bias: Low Imprecise results

Moderate usefulness

Treggiari et al. 200932

RCT Selection bias: Low Performance bias: Low Attrition bias: Low Detection bias: Unclear

Reporting bias: Low Imprecise results

Moderate usefulness

Ethier et al. 201130 RCT Selection bias: Low bias: : Low s: Low

Reporting bias: Moderate

s v o s a p ( o d

g t

Performance Attrition bias Detection bia

tudy was designed to examine the effect of protocolised sedation s. protocolised sedation and daily sedative interruption on mem- ries, with findings indicating that recall was not correlated with edation scores or doses of sedation received.30 Similarly, Kress nd colleagues31 found no difference in the number of usual care atients vs. sedation interruption patients who recalled being in 68% vs. 69%, p = 1.0), or waking in (26% vs. 0%, p = 0.06), ICU. Both f these studies were small (n = 21 and 32), with limited ability to

etect meaningful differences.

The relatively limited evidence regarding sedation and anal- esia in ICU patients suggests that these therapies contribute to he patients’ ability to recall ICU events and to the development

Unclear Low Imprecise results

usefulness

of delusional memories and amnesia of ICU stay. However, the evidence is conflicting and inconclusive. Larger studies explor- ing different aspects of sedation and their association with the development of memories of the ICU during recovery are needed.

3.2. Other medications

The theory that the exogenous administration of stress doses

of corticosteroids provides a protective effect against the devel- opment of traumatic memories was tested in two small RCTs of cardiac surgical ICU patients.24,35 Although some other benefits such as shorter ICU LOS and improved quality of life scores were

1 an Cri

s w R d t t t p r a o 3 w g r t s b t p

3

p t a o i m t t m v a i m o

3

i r M t Q i w m t s o f a

4

m S i b d t w

2 L.M. Aitken et al. / Australi

een in those receiving corticosteroids, no significant difference as found in the incidence of traumatic memories in either of these CTs. In contrast, a significant association between corticoids and elusional memories (OR 10.2 95% CI 1.11–93.0, p = 0.04) was iden- ified in a small (n = 161) retrospective cohort study.4 Although his finding was statistically significant, it is important to note he limitations of the study design and that only eight out of 161 atients received corticoids during ICU, with six of these patients eporting delusional memories.4 In this same study a significant ssociation between delusional memories and the administration f intravenous 50% glucose to treat hypoglycaemia (OR 15.5, 95% CI .19–66.4) was identified. To aid understanding of these results it ould have been beneficial if data regarding the severity of hypo-

lycaemia were presented since one could speculate that the real isk factor for delusional memories was hypoglycaemia instead of he administration of 50% glucose. Another point to consider is the urprisingly high proportion (35%) of participants who received oluses of glucose during ICU treatment, which could reflect selec- ion bias with participants not being representative of the usual ICU opulation.

.3. Duration of mechanical ventilation and ICU stay

The presence or duration of mechanical ventilation has also been roposed as influencing memories after ICU, particularly in regard o the development of traumatic or distressing memories or the bsence of memories. The duration of mechanical ventilation is ften related to the length of ICU stay which may be used as an ndirect marker of this treatment. Approximately half of the 206

echanically ventilated general ICU patients who had memories of he ICU recalled discomfort associated with the endotracheal tube hat were bothersome and those who were bothered by stressful

emories of the ICU had longer ICU LOS.18 In contrast, although entilator support was linked to delusional memories on univariate nalysis in 239 trauma patients, it did not retain statistical signif- cance when incorporated into multivariable analysis.34 Similarly,

echanical ventilation was not associated with factual, sensational r emotional memories in 152 ICU patients.1

.4. Measures and methods used to assess patients’ memories

Memories were measured using a variety of instruments ncluding structured interviews,1,31,36 the ICU stressful expe- ience questionnaire (ICU-SEQ),2,3,18 ICU memory tool (ICU- T)2,3,17,18,34, Posttraumatic Stress Symptoms 10-Question Inven-

ory PTSS-1023,35 and author-developed surveys (Patients Recall uestionnaire30 and ICU amnesia tool)7 or questions.31 A mod-

fied version of the ICU-MT (with no validation of modification) as used in one study.4 In another study no specific instru- ent to assess patients’ memories was used, but two items of

he Post-Traumatic Stress Disorder Checklist (instrument to assess ymptoms of posttraumatic stress) that address memories.32 Mem- ries were assessed at varied time points in the included studies, rom 72 h30 to 5 years4, with the most common follow-up being pproximately six months after discharge (Table 1).

. Discussion

The relationship between intensive care interventions and emories of ICU after discharge was examined in this review.

edation practice was the most common intensive care treatment nvestigated in relation to the development of memories of ICU,

ut the evidence was inconsistent for the elements of care (e.g. eep vs. light sedation, different sedative medications, daily seda- ion interruption). Deep sedation during ICU treatment frequently as associated with amnesia and delusional memories while light

tical Care 29 (2016) 5–14

sedation was associated with a greater risk of perceiving stress- ful experiences more bothersome.2,3,18,32 Despite these identified associations, no such association was reported in one study.7 This relationship between level of sedation and memories after ICU is particularly important to understand given the move towards lighter sedation over the past decade.26

When considering the specific sedative agents, benzodiazip- ines, including midazolam, were associated with the development of delusional or hallucination-like memories of ICU in some settings.4,17 This relationship with delusional memories warrants further exploration since the available evidence is small and incon- sistent. Although this relationship was found in two included studies, it was not identified in others.1,34 Since these four stud- ies differ in essential aspects of their design (e.g. prospective vs. retrospective; cardiac vs. trauma vs. general ICU patients), the com- parison between them might not be appropriate and therefore the interpretation of the evidence in regard to benzodiazipines is incon- clusive.

Different sedation strategies such as using sedation protocols, daily sedation interruption and various sedative agents such as midazolam, isoflurane, morphine and propofol were tested to eluci- date their association with memories of ICU.1,17,30,31 No particular strategy was found to be better or worse than others. Neverthe- less, these studies had relatively small sample sizes with restricted ability to determine significant differences. Future studies testing different sedation strategies should incorporate larger sample sizes in their design so as to be able to detect significant effects.

Interventions other than sedation that have been examined in relation to ICU memories were corticosteroids and intravenous 50% glucose to treat hypoglycaemia. The evidence on the associ- ation between stress doses of corticosteroids and memories of ICU is limited. As slightly different aspects of memories of ICU were explored in these studies, the comparison between them is diffi- cult. In two RCTs the factor explored was the incidence of traumatic memories (memories of pain, nightmares, anxiety and difficulty breathing) compared with the presence of delusional memories in one retrospective study.4,24,35 Delusional memories were associ- ated with the administration of corticosteroids, but no association between hydrocortisone and the incidence of traumatic memories of anxiety, pain, nightmares and difficulty breathing was found in the RCTs.

Intravenous 50% glucose to treat hypoglycaemia was associated with delusional memories.4 Unfortunately, information regarding the severity of hypoglycaemia was not presented and it could be speculated that 50% glucose might be a confounder and that hypoglycaemia might have been the real risk factor for delusional memories. In addition, the high proportion of participants treated with this medication suggests selection bias.

No relationship between mechanical ventilation and delusional, factual, sensational or emotional memories was found in the literature.1,34 Despite the evidence being limited to two studies, the lack of relationship is consistent.

This review is limited by the nature of the question that guided the process; only those studies that measured an association between ICU interventions and memories were considered, result- ing in the review being limited to studies designed to measure variables quantitatively. Also to note, a range of interventions that have the potential to influence memory have not been investigated, for example early mobilisation and use of alternative sedative agents such as dexmedetomidine. Further, scales that assess either memory or perceptions of the intensive care experience generally have not undergone rigorous psychometric testing thus limiting

reliability and validity of findings. A number of scales exist and this also makes it difficult to extrapolate consistent findings with some assessing memories of specific events or treatments and oth- ers perceptions of the experience. What however does seem clear

an Cri

i a c a a t e f W t s d

h h c a m o c

4

t s o i D r s s p n

i b h p i e d d

5

d A g e f i h

A

a

Q 7

1

1

1

1

1

1

1

1

1

1

2

2

2

2

2

2

2

2

L.M. Aitken et al. / Australi

s that patients can be distressed by their memories whether these re factual or delusional,11,37,38 and that these memories have been onsistently linked to poorer outcomes.13,18,39 The strength of that ssociation has still to be established. Patients who have greater wareness may report troublesome discomforts of thirst, having ubes and being unable to communicate1,37 whereas patients who xperience delusions often find these persecutory and there are requent reports of staff trying to ‘kill’ patients or do harm to them.

hether these differences in perceptions or memories are related o depth of sedation is not clear, although the evidence that deep edation leads to limited recall of ICU and increased incidence of elusional memories is reasonably consistent.2,3,18,32

Participants in the included studies were enrolled during their ospital admission, with the exception of one retrospective study,4

owever the timing of follow up varied from prior to ICU dis- harge to 5 years later, with the most common follow-up being pproximately six months after discharge. Given the highly variable ethods of assessing memory it is not possible to assess the effect

f these differences, however it is highly likely that it influences the ontent and clarity of recall.

.1. Implications for practice and research

Although there is currently limited and inconsistent evidence, he influence of sedation on memories has moderate support. Deep edation is linked to limited recall of ICU and increased incidence f delusional memories.2,3,32 This suggests that strategies to min- mise sedation should continue to be developed and implemented. espite this broad principle, there is currently conflicting evidence

egarding the role of different sedation strategies such as daily edation interruption30,31 or the benefit or disadvantage of specific edative agents4,17 and additional research involving larger sam- le sizes and effective control of related interventions is urgently eeded.

Further, scales that assess either memory or perceptions of the ntensive care experience generally assume that patients who recall eing ‘attacked’ or have ‘people trying to hurt me’ are delusional owever this might not be the case. Instead patients may be inter- reting behaviours in different, and individual, ways. This variation

n experience should be considered as we research and implement volving strategies such as patients being more alert and oriented uring ICU admissions – this experience may affect each patient ifferently.

. Conclusion

Identification of elements of ICU treatment that affect memories uring recovery has the potential to influence how care is delivered. spects of care that have been examined include sedation and anal- esia, other medications and mechanical ventilation. Although the vidence was inconsistent, and the numbers of participants was requently small, it appears that some aspects of treatment may nfluence the absence of memory or development of delusions and allucinations.

cknowledgements

All authors have approved the final article and acknowledge that ll those entitled to authorship are listed as authors.

To answer the Continuing Professional Development uestions - go to page 12 http://dx.doi.org/10.1016/S1036- 314(16)00010-2.

2

tical Care 29 (2016) 5–14 13

References

1. Capuzzo M, Pinamonti A, Cingolani E, Grassi L, Bianconi M, Contu P, et al. Anal- gesia, sedation, and memory of intensive care. J Crit Care 2001;16:83–9.

2. Samuelson K, Lundberg D, Fridlund B. Memory in relation to depth of seda- tion in adult mechanically ventilated intensive care patients. Intensive Care Med 2006;32:660–7.

3. Samuelson KA, Lundberg D, Fridlund B. Light vs. heavy sedation during mechan- ical ventilation after oesophagectomy – a pilot experimental study focusing on memory. Acta Anaesthesiol Scand 2008;52:1116–23.

4. Pierce JMT, Allenby-Smith O, Goddard J. Delusional memories following car- diac surgery and prolonged intensive care: a retrospective survey and case note review. Brit J Cardiol 2004;11:462–5.

5. Jones C, Humphris G, Griffiths RD. Preliminary validation of the ICUM tool: a tool for assessing memory of the intensive care experience. Clin Intensive Care 2000;11:251–5.

6. Schelling G, Stoll C, Haller M, Briegel J, Manert W, Hummel T, et al. Health- related quality of life and posttraumatic stress disorder in survivors of the acute respiratory distress syndrome. Crit Care Med 1998;26:651–9.

7. Weinert CR, Sprenkle M. Post-ICU consequences of patient wakefulness and sedative exposure during mechanical ventilation. Intensive Care Med 2008;34:82–90.

8. Needham DM, Davidson J, Cohen H, Hopkins RO, Weinert C, Wunsch H, et al. Improving long-term outcomes after discharge from intensive care unit: report from a stakeholders’ conference. Crit Care Med 2012;40:502–9.

9. Davydow DS, Zatzick D, Hough CL, Katon WJ. In-hospital acute stress symptoms are associated with impairment in cognition 1 year after intensive care unit admission. Ann Am Thorac Soc 2013;10:450–7.

0. Karnatovskaia LV, Johnson MM, Benzo RP, Gajic O. The spectrum of psychocog- nitive morbidity in the critically ill: a review of the literature and call for improvement. J Crit Care 2015;30:130–7.

1. Jones C, Griffiths RD, Humphris G, Skirrow PM. Memory, delusions, and the development of acute posttraumatic stress disorder-related symptoms after intensive care. Crit Care Med 2001;29:573–80.

2. Jones C, Skirrow P, Griffiths RD, Humphris GH, Ingleby S, Eddleston J, et al. Rehabilitation after critical illness: a randomized, controlled trial. Crit Care Med 2003;31:2456–61.

3. Ringdal M, Plos K, Lundberg D, Johansson L, Bergbom I. Outcome after injury: memories, health-related quality of life, anxiety, and symptoms of depression after intensive care. J Trauma 2009;66:1226–33.

4. Kiekkas P, Theodorakopoulou G, Spyratos F, Baltopoulos GI. Psychological dis- tress and delusional memories after critical care: a literature review. Int Nurs Rev 2010;57:288–96.

5. Ringdal M, Plos K, Ortenwall P, Bergbom I. Memories and health-related quality of life after intensive care: a follow-up study. Crit Care Med 2010;38:38–44.

6. Deja M, Denke C, Weber-Carstens S, Schroder J, Pille CE, Hokema F, et al. Social support during intensive care unit stay might improve mental impairment and consequently health-related quality of life in survivors of severe acute respira- tory distress syndrome. Crit Care 2006;10:R147.

7. Sackey PV, Martling CR, Carlsward C, Sundin O, Radell PJ. Short- and long-term follow-up of intensive care unit patients after sedation with isoflurane and midazolam – a pilot study. Crit Care Med 2008;36:801–6.

8. Samuelson KA, Lundberg D, Fridlund B. Stressful memories and psychological distress in adult mechanically ventilated intensive care patients – a 2-month follow-up study. Acta Anaesthesiol Scand 2007;51:671–8.

9. Hauer D, Weis F, Krauseneck T, Vogeser M, Schelling G, Roozendaal B. Trau- matic memories, post-traumatic stress disorder and serum cortisol levels in long-term survivors of the acute respiratory distress syndrome. Brain Res 2009;1293:114–20.

0. Granja C, Gomes E, Amaro A, Ribeiro O, Jones C, Carneiro A, et al. Understand- ing posttraumatic stress disorder-related symptoms after critical care: the early illness amnesia hypothesis. Crit Care Med 2008;36:2801–9.

1. Larson MJ, Weaver LK, Hopkins RO. Cognitive sequelae in acute respiratory dis- tress syndrome patients with and without recall of the intensive care unit. J Int Neuropsychol Soc 2007;13:595–605.

2. de Quervain DJ, Roozendaal B, Nitsch RM, McGaugh JL, Hock C. Acute cortisone administration impairs retrieval of long-term declarative memory in humans. Nat Neurosci 2000;3:313–4.

3. Schelling G, Kilger E, Roozendaal B, de Quervain DJ, Briegel J, Dagge A, et al. Stress doses of hydrocortisone, traumatic memories, and symptoms of post- traumatic stress disorder in patients after cardiac surgery: a randomized study. Biol Psychiatry 2004;55:627–33.

4. Schelling G, Roozendaal B, De Quervain DJ. Can posttraumatic stress disorder be prevented with glucocorticoids. Ann NY Acad Sci 2004;1032:158–66.

5. Jones C, Griffiths RD, Humphris G. Disturbed memory and amnesia related to intensive care. Memory 2000;8:79–94.

6. Barr J, Fraser GL, Puntillo K, Ely EW, Gelinas C, Dasta JF, et al. Clinical practice guidelines for the management of pain, agitation, and delirium in adult patients in the intensive care unit. Crit Care Med 2013;41:263–306.

7. Green BN, Johnson CD, Adams A. Writing narrative literature reviews for peer- reviewed journals: secrets of the trade. J Chiropract Med 2006;5:101–17.

8. Clinical Appraisal Skills Programme (CASP), Oxford UK, http://www.casp-uk. net/Date [accessed 19.10.15].

1 an Cri

2

3

3

3

3

3

3

3

3

38. Rundshagen I, Schnabel K, Wegner C, am Esch S. Incidence of recall, nightmares, and hallucinations during analgosedation in intensive care. Intensive Care Med 2002;28:38–43.

4 L.M. Aitken et al. / Australi

9. Moher D, Liberati A, Tetzlaff J, Altman DG. Preferred reporting items for sys- tematic reviews and meta-analyses: the PRISMA statement. J Clin Epidemiol 2009;62:1006–12.

0. Ethier C, Burry L, Martinez-Motta C, Tirgari S, Jiang D, McDonald E, et al. Recall of intensive care unit stay in patients managed with a sedation protocol or a sedation protocol with daily sedative interruption: a pilot study. J Crit Care 2011;26:127–32.

1. Kress JP, Gehlbach B, Lacy M, Pliskin N, Pohlman AS, Hall JB. The long-term psy- chological effects of daily sedative interruption on critically ill patients. Am J Respir Crit Care Med 2003;168:1457–61.

2. Treggiari MM, Romand JA, Yanez ND, Deem SA, Goldberg J, Hudson L, et al. Ran- domized trial of light versus deep sedation on mental health after critical illness. Crit Care Med 2009;37:2527–34.

3. Henderson A, Dipplesman J, Miller J. Failure of intravenous low dose midazolam

to influence memory recall in drug paralysed post-operative patients sedated with papaveretum. Aust Crit Care 1994;7:22–4.

4. Ringdal M, Johansson L, Lundberg D, Bergbom I. Delusional memories from the intensive care unit – experienced by patients with physical trauma. Intensive Crit Care Nurs 2006;22:346–54.

3

tical Care 29 (2016) 5–14

5. Weis F, Kilger E, Roozendaal B, de Quervain DJ, Lamm P, Schmidt M, et al. Stress doses of hydrocortisone reduce chronic stress symptoms and improve health- related quality of life in high-risk patients after cardiac surgery: a randomized study. J Thorac Cardiovasc Surg 2006;131:277–82.

6. Wagner BK, Zavotsky KE, Sweeney JB, Palmeri BA, Hammond JS. Patient recall of therapeutic paralysis in a surgical critical care unit. Pharmacotherapy 1998;18:358–63.

7. Rotondi AJ, Chelluri L, Sirio C, Mendelsohn A, Schulz R, Belle S, et al. Patients’ recollections of stressful experiences while receiving prolonged mechanical ventilation in an intensive care unit. Crit Care Med 2002;30: 746–52.

9. Rattray JE, Johnston M, Wildsmith JA. Predictors of emotional outcomes of inten- sive care. Anaesthesia 2005;60:1085–92.