- Email: [email protected]
Alterations in cognitive outcome between 3 and 12 months in survivors of out-of-hospital cardiac arrest
Resuscitation 105 (2016) 92–99
Contents lists available at ScienceDirect
Resuscitation journal homepage: www.elsevier.com/locate/resuscitation
Clinical paper
Alterations in cognitive outcome between 3 and 12 months in survivors of out-of-hospital cardiac arrest Marte Ørbo a,∗ , Per M. Aslaksen b,c , Kristina Larsby d , Christoph Schäfer a , Pål M. Tande d , Audny Anke a,f a
Department of Rehabilitation, Division of Neurosciences, Orthopedics and Rehabilitation Services, University Hospital of North Norway, Norway Department of Psychology, Faculty of Health Sciences, University of Tromsø, Norway c Department of Child and Adolescent Psychiatry, Division of Child and Adolescent Health, The Regional Unit for Eating Disorders, University Hospital of North Norway, Norway d Department of Cardiology, Division of Cardiothoracic and Respiratory Medicine, University Hospital of North Norway, Norway f Department of Clinical Medicine, Faculty of Health Sciences, University of Tromsø, Norway b
a r t i c l e
i n f o
Article history: Received 25 September 2015 Received in revised form 17 April 2016 Accepted 15 May 2016 Keywords: Cardiac arrest Cognitive Rehabilitation Recovery Neuropsychology
a b s t r a c t Objectives: To prospectively investigate cognitive recovery from 3 to 12 months after resuscitation from out-of-hospital cardiac arrest (OHCA) and the associations between cognitive performance at 3 months and health-related quality of life (HRQL), psychological distress and work status after 12 months. Methods: At both assessments, neuropsychological tests were used to measure aspects of general mental ability, verbal and visual memory, psychomotor speed and executive function. The Short Form-36 (SF-36) was used to measure mental and physical HRQL, and the Hospital Anxiety and Depression Scale (HADS) to assess psychological distress. Results: 33 survivors completed both exams (31 males, mean age 58.6 years, SD = 13). The OHCAs were witnessed and due to cardiac origins. Nine patients were awake at admission to the hospital. Longer coma duration was associated with poorer cognitive results. Memory impairments were the most common symptom. The mean changes and effect sizes indicated minor improvements in cognitive performance from 3 to 12 months (Hedges g ≤ .26). Reliable change indices for an individual’s results further confirmed the stability of the group statistics. The HADS scores showed increased depressive symptoms, and mental HRQL was reduced from 3 to 12 months. Higher reports of psychological distress were related to worse HRQL. Work participation increased. Better cognitive results at 3 months were correlated with better HRQL and return to work at 12 months. Conclusions: The current data describe stability in results from 3 to 12 months. A worse cognitive performance at 3 months and higher reports of psychological distress were associated with lower HRQL. © 2016 Elsevier Ireland Ltd. All rights reserved.
Introduction Cognitive dysfunction is estimated to occur in 30–50% of survivors of out-of-hospital cardiac arrest (OHCA) when measured with objective, performance-based neuropsychological tests in the first year after resuscitation.1–3 Most frequently identified are problems with memory functions.1–4 In addition to the hypoxic-ischemia caused by the circulatory arrest, other risk factors for cognitive decline exist in this patient group and include
∗ Corresponding author at: Department of Rehabilitation, Division of Neurosciences, Orthopedics and Rehabilitation Services, University Hospital of North Norway, 9038 Tromsø, Norway. E-mail address: [email protected] (M. Ørbo). http://dx.doi.org/10.1016/j.resuscitation.2016.05.017 0300-9572/© 2016 Elsevier Ireland Ltd. All rights reserved.
pre-existing cardiovascular burden and general critical illness.1,5 Despite the existence of multiple risk factors for cognitive problems, cognitive decline is not regularly assessed and may go unrecognized in post-cardiac arrest care and follow-up.3,6,7 Previous studies of cognitive outcomes have most often relied on a single time point of measurement.3 Few previous prospective studies have assessed improvement in cognitive performance based on repeated neuropsychological testing of the same survivors over time.3,8–12 It is generally suggested that cognitive recovery plateaus in the first 3 months after resuscitation and that early cognitive assessments approximate longer-term outcomes.3 The aim of the present study was to assess improvement in cognitive functioning 3–12 months after resuscitation. Additionally, this study aimed to explore if worse cognitive functioning detected at 3 months anticipated worse functional prognosis in terms of poorer health related
M. Ørbo et al. / Resuscitation 105 (2016) 92–99
quality of life (HRQL), more psychological distress or not returning to work 1 year after an OHCA. Based on the results from previous studies, we expected little cognitive improvement beyond 3 months3,10 but believed that lower cognitive performance could be associated with lower HRQL after 1 year.6,10,13 We further expected that higher reported levels of psychological distress would negatively influence HRQL.6,14 Methods The study was approved by the Regional Committee for Medical Research Ethics in North-Norway, institutional protocol number 2009/1395. Written consent was obtained from all participants. Participants We aimed to prospectively include all survivors of nontraumatic, normothermic OHCA of presumed cardiac origin, aged 18–85 years, who were discharged from the cardiac ward at the University hospital of North Norway between August 2010 and September 2013 and who were able to undergo a neuropsychological assessment 3 months after resuscitation. The hospital is a tertiary care center with an advanced cardiac ward serving smaller hospitals in a regional manner. According to hospital records, 197 survivors were treated for OHCA of any cause in the study period. Patients who died before or during hospital admission were not registered. Of these OHCAs, 129 were of cardiac origin. Participants had to be fluent Norwegian speakers, be living independently prior to cardiac arrest and have no history of psychiatric or neurologic disease or significant ongoing somatic disease, substance abuse or alcoholism. Fig. 1 shows a flow-chart starting with the total number of OHCA patients discharged alive from the cardiac ward during the study period who had an OHCA due to a presumed cardiac origin; the reasons for exclusion and loss to follow-up at both time points for assessment are also displayed. Procedure Survivors were invited to the hospital for an individual, outpatient, face-to-face follow-up assessment at 3 months (T1) and at 12 months (T2) post-OHCA that included performance-based neuropsychological tests and 2 self-report questionnaires: the Hospital Anxiety and Depression Scale (HADS)15 and the Short Form (36) Health Survey (SF-36) version 1.2.16 Experienced neuropsychologists administered and scored all tests according to standardized procedures. Tests and questionnaires were administered in one session at each time point. Age, education history, living situation and working status prior to cardiac arrest and at both follow-up time points were recorded from the patient interview. Information about their medical history and treatment variables were recorded from participants’ medical journals at the hospital. Description of the outcome measures Neuropsychological tests Details of the individual neuropsychological tests17–22 are described in Supplementary Table 1. We selected tests that were representative of the cognitive functions described as being the most frequently impaired in OHCA survivors: declarative memory, executive function and psychomotor function.2,3,11,13,23 The study was performed in a clinical setting, and thus only tests that are commonly applied in clinical practice were used. The tests selected have been widely used internationally and have been standardized to the Norwegian setting, with psychometric properties such as construct validity and test-retest reliability that have empirically been shown to be satisfactory to excellent. Published normative data from large groups of healthy individuals were used to
93
score each test correcting for demographic variables.17–22 The standardized test results were transformed to Z-scores: Patient score − normative mean score Normative standard deviation of the test The tests were grouped into five cognitive domain composite scores based on construct validity. A mean Z-score for each domain was computed to both reduce the number of comparisons in the statistical analyses and to reduce the influence of possible idiosyncratic results that may have occurred by chance on single tests. A cut-off of ≤−1.5 SD from the normative mean was used to characterize abnormal performance on each cognitive composite score. Self-report questionnaires The HADS is a brief self-report questionnaire that consists of 14 questions; 7 questions assess anxiety symptoms and 7 assess symptoms of depression (score 0–3 for each question), each scale having a maximum score of 21.24 Scores at or above eight points on either the depression or anxiety scale are suggested to be clinically relevant.14,24 The questionnaire is frequently used in studies of cardiac arrest survivors and has been shown to have sound psychometric properties.14 HRQL was measured with the Medical Outcomes Study 36Item Short Form Health Survey (SF-36 version 1.2), a generic HRQL measure that consists of 2 summary scales, a physical component scale (PCS) and a mental component scale (MCS), which are divided into 8 subscales: physical functioning (PF), physical role (RP), bodily pain (BP), general health (GH), mental health (MH), emotional role (RE), social function (SF) and vitality (VT).16,25,26 The questionnaire asks patients about their perceptions of how their health status has interfered with their psychological, social and physical functioning in the previous 4 weeks. The participants’ data were compared with age- and gender-corrected normative data from the general Norwegian population.25 An online calculator for norm-based data scoring (http://www.sf36.org/nbscalc/index. shtlml) was applied for each survey, and the results were provided in normalized T-scores. When standardizing the scores according to the T-score distribution, the means were 50 and the standard deviations were 10 across all summary scales and subscales on the SF-36 in the comparator group. Thus, a direct comparison between the participants’ T-scores and the age- and gender-corrected normative data in the general Norwegian population could be viewed in a single graph. Fig. 2 shows the standardized SF-36 results from the 33 participants at both time points of assessment. T < 40 on any of the scales is suggested to represent HRQL below the normal range.26 Statistical analyses Statistical analyses were performed in IBM SPSS (version 22). Missing data at T1 were found in three participants on the Rey Complex Figure and in two participants on the Grooved Pegboard tests. At T2, missing data were found for one participant on the Grooved Pegboard test and in two participants on the WordFluency Test. In the HADS data, two participants had missing data. All missing data were replaced by the series mean. Before performing inferential statistics, all continuous variables were examined for normality by the Shapiro–Wilk test and by visual inspection of the Q–Q plots and box-plots of the distributions. Continuous variables were analyzed by parametric tests, as none of the continuous data were significantly different from a normal distribution. Statistical group comparisons were performed by one-way analysis of variance (ANOVA). To test whether the mean of the continuous data differed from normative data, one-sample t-tests were
94
M. Ørbo et al. / Resuscitation 105 (2016) 92–99
Discharged alive after OHCA due to cardiac cause August 2010 – September 2013 Excluded due to predefined criteria N = 129
Age above 85 years, n = 2 Severe anoxic brain injury, n = 8 Tourists/living abroad, n = 10 Non-fluent in Norwegian language, n = 1 Other severe somatic illness, n = 7 Psychiatric disorder / neurological disease / learning disabilities / drug/alcohol abuse, n = 18
Died before 3 months after OHCA n=4 Eligible for the study N = 79
Lost to follow-up Transferred to further treatment and rehabilitation/mild anoxic injury suspected, n = 6 Discharged to own home – no brain injury suspected and not asked to participate, n = 13
Refused to participate Reasons: Long travelling distance, n = 2 Included at 3 months
The examination felt unnecessary due to selfreported good functioning, n = 10
N = 45
Did not want to participate in research, n = 2 No show, n = 1
Participants lost from 3 to 12 months Reasons: Felt no need for a second evaluation due to self-reported good functioning, n = 3 Felt no use of a second evaluation, n = 1 Not asked to participate due to time restrictions in the data collection, n = 5. Logistics, n = 2 Somatic illness, n = 1 Included at 12 months N=33
Fig. 1. Overview of enrollment and exclusion. In addition to the predefined exclusion criteria shown in this figure, previous cardiac arrest and cardiac surgery during the year prior to OHCA were set as predefined reasons for exclusion. None of the patients included had experienced previous cardiac arrest nor had they received cardiac surgery in the year prior to OHCA.4
employed. Group comparisons with several dependent variables were performed by multivariate ANOVA. The analyses of changes from T1 to T2 in neuropsychological composite scores, HADS scales scores and SF-36 summary component scores were first performed with paired-sample t-tests. Then, the Holm-Bonferroni method was used to adjust the p-values
for multiple comparisons in the analyses assessing change from T1 to T2. In all other analyses, a p-value < .05 was considered significant. Hedges’ g was calculated to display the effect sizes of the mean change in continuous variables from 3 to 12 months post-OHCA. Effects were considered small when g was <.5, medium for g < .5–.8,
M. Ørbo et al. / Resuscitation 105 (2016) 92–99
95
Fig. 2. Comparisons between the subscale and summary component scores of the SF-36 at T1 and T2. The results from T1 are displayed first in each pair. Higher scores indicate better HRQL. SF-36 scores are shown in normed T-scores, with a mean = 50 and SD (standard deviation) = 10. Error bars denote SD. The solid line indicates the normative mean, and the two dotted lines indicate ±1 SD from the normative mean. PF = physical functioning. RP = role physical. BP = body pain. GH = general health. VT = vitality. SF = social functioning. RE = role emotional. MH = mental health. PCS = physical component summary. MCS = mental component summary. N = 33. * indicates significant difference (p < .05) tested by pairwise t-tests, uncorrected for multiple comparisons.
and large for g > .8.27 The change in frequency data (work status) was examined by the related samples McNemar’s change test. Reliable change indexes (RCI)28 were calculated for the neuropsychological domain composites, SF-36 summary components (PCS and MCS) and the HADS anxiety and depression scales according to the following formula:
T2 score − T1 score
2SDT12 (1 − correlation between T1 and T2)
An RCI-value indicating a change of ±1.645 SD was considered a statistically reliable change.29 Correlations between the neuropsychological composites and self-report measures were performed by two-tailed Pearson correlations. Results Sample characteristics The demographic, medical and resuscitation characteristics are displayed in Table 1. One-way ANOVA detected no group differences in duration of coma, age, years of education, ICD, hypothermia, time to ROSC or employment status, cognitive tests at T1 or self-report questionnaires at T1 (all Fs < 2.9, all p values > 0.1) between participants having neuropsychological data at both time points (N = 33) and in those who only had data from T1 (N = 12). The participants had variable coma durations, and nine were awake upon admission to the hospital. A multivariate ANOVA showed no significant differences in the cognitive composite scores, HADS scores or SF-36 MCS or PCS scores between participants who were awake at admission and those who were unconscious (all Fs < 3.95). Data for the participants who were awake at admission compared to those who were unconscious at admission are shown in Supplementary Table 2. Only one of the patients awake at admission scored in the cognitively impaired range. A longer coma duration was significantly associated with poorer performance in all cognitive domains at T1 and T2 (all p values <.01 adjusted for multiple comparisons, all Fs > 5.36) but not with verbal memory at T2.
Descriptive results at T1 and T2 Table 2 presents the mean, SD and range for the five cognitive composite scores, the HADS depression and anxiety scores and the SF-36 summary component scores (PCS and MCS) at both T1 and T2. Additionally, descriptive information of all SF-36 subscale scores are provided in Fig. 2. Table 3 shows the percentage of abnormal scores according to the defined cut-off points for the HADS scales, SF-36 summary component scores and cognitive composite scores. Work status is shown in Table 1. Assessment of changes in outcome variables from T1 to T2 Table 2 shows the results of the analyses of statistical changes in the cognitive composite, SF-36 summary component and HADS scores between T1 and T2 in terms of t-tests and effects sizes. For cognitive outcomes, the paired t-tests suggested significant improvement in two cognitive domains: visual memory (t = −2.46, p = .02) and executive function (t = −2.31, p = .03). The effect sizes were small (Hedges g ≤ .26). When lowering the significance level according to the Holm-Bonferroni adjustment for multiple comparisons, a significant improvement was shown in the visual memory domain only (p < .025). A significant decline was found from T1 to T2 in the MCS scale score on the SF-36, indicating a worsening of mental HRQL. When examining changes in the 8 different SF-36 subscales that constitute the summary components, no significant statistical change was revealed between T1 and T2 (all ps > .063) by two-tailed t-tests (Fig. 2). Table 2 further shows a significant increase in HADS depression but not anxiety scores from T1 to T2. The number of active workers significantly increased from T1 to T2 (McNemar = 4.9, df = 1, p = .021). Table 3 shows the RCI indexes for cognitive composite scores, the HADS scores and SF-36 summary component scores. The additional information on cognitive outcomes provided was that eight persons showed reliable improvement in one cognitive domain, and three persons showed improvement in two domains, suggesting some reliable improvement in 33% of the participants.
96
M. Ørbo et al. / Resuscitation 105 (2016) 92–99
Table 1 Demographical and medical characteristics of the sample. Demographics
Number (%)
Male Female Age Married/partner Years of education Active workers before OHCA Active workers after 3 months Active workers after 12 months Living in own home at 3 months Living in own home at 12 months
31 (94) 2 (6)
Mean (SD)
Min–Max
58.6 (13)
18–82
12 (4)
5–21
16 (12)
1–60
42.9 (57)
2 min–192 h
4.8 (4.1) 11.4 (6.2)
1–14 5–36
28 (85) 18 (55) 2 (6) 11 (33) 32 (97) 33 (100)
Medical history No cardiovascular history Hypertension Myocardial infarction Heart failure Ischemia Diabetes mellitus
19 (58) 5 (15) 7 (21) 1 (3) 1 (3) 2 (6)
OHCA and treatment characteristics First rhythm VF AF Asystole Time to ROSC – minutes Cause of arrest Infarction Rhythm Unknown Awake at admission Duration of coma in hours Hypothermia PCI CABG ICD Medication
31 (94) 1 (3) 1 (3)
27 (82) 4 (12) 2 (6) 9 (27) Yes 14 (42)a Yes 26 (79) Yes 2 (6) Yes 10 (30) 23 (69)
Time variables Length of hospital stay Intensive care unit Cardiac ward Rehabilitation Days from OHCA to T1 Days from OHCA to T2
110 (15) 371 (33)
79–122 312–419
Overview of participant’s demographic, resuscitation and medical variables (N = 33). All had witnessed OHCA with bystander cardiopulmonary resuscitation (CPR) initiated. VF = ventricular fibrillation. AF = atrial fibrillation. ROSC = return of spontaneous circulation. PCI = percutaneous coronary intervention. ICD = implanted cardioverter defibrillator. CABG = Coronary Artery Bypass Grafting. T1 = assessment after 3 months. T2 = assessment after 12 months. a Reasons for not receiving therapeutic hypothermia were (N = 19): awake at admission (9), immediate CABG after admission (1), renal complications (1), uncertain state of consciousness at admission (2), pneumonia (1), aspiration pneumonia (1), long transportation distance to hospital (2), sepsis (1), procedural complication, hypothermia terminated before target temperature (1). Table 2 Assessment of change in outcomes from T1 to T2 (N = 33). Mean T1 (SD)
Range T1
Mean T2 (SD)
Range T2
r (T1–T2)
t
p
Cognitive composite Visual memory composite Verbal memory composite Psychomotor composite Executive composite WASI IQa
Hedges g
CI (95%) for g
−.60 (1.51) −.49 (1.13) −.01 (.64) −.13 (1.15) 98 (15.43)
−3 to 2.3 −3.4 to 1.8 −1.9 to 1.1 −3 to 1.3 60–126
.39 (1.57) −.27 (1.28) .15 (.53) .12 (.87) 99 (15.33)
−3 to 3 −3.5 to 1.6 −1.4 to 1.3 −2.7 to 1.7 70–127
.84** .80** .72** .83** .88**
−2.46 −1.65 −1.96 −2.31 −.78
.02† .11 .06 .03 .44
.26 .18 .25 .25 .07
−.25 to .76 −.30 to .66 −.24 to .74 −.24 to .75 −3.55 to 3.68
Self-report measures HADS anxiety HADS depression SF-36 PCSb SF-36 MCSb
2.94 (2.92) 1.7 (1.89) 44.10 (8.55) 51.85 (9.57)
0–11 0–5 29.7–56.9 29–63
3.78 (3.8) 2.56 (3.24) 45.82 (8.04) 47.71 (11.82)
0–16 0–12 29.7–57 18.4–63.4
.62** .80** .92** .70**
−1.71 −2.5 −1.58 2.60
.09 .02† .12 .02†
.27 .35 .21 −.38
−.55 to 1.09 −.30 to .99 −1.80 to 2.21 −2.98 to 2.21
Paired samples t-tests and effects sizes of the change (Hedges g). Cognitive composites are displayed as Z-scores (mean = 0, SD = 1). Higher Z-scores indicate better performance. r = Pearson correlation. ** p < .001. † Significant change after correction for multiple comparisons by the Holm-Bonferroni procedure, adjusting for correlation between the measures and degrees of freedom. Adjustments for multiple comparisons were performed separately for cognitive (p < .025) and self-report measures (p < .035). Mean correlation between the cognitive composites was r = .56, and mean correlation between the self-report measures was r = .76. p values indicate difference in paired-sample t-tests. CI = Confidence interval. a WASI IQ is displayed as Wechsler scores, normative mean = 100, SD = 15. Higher scores indicate better performance. HADS = Hospital Anxiety Depression Scale, higher scores on both subscales indicates more symptoms. SF-36 = Short Form 36. b SF-36 is displayed in T-scores (normative mean = 50, SD = 10), higher scores indicates better HRQL. PCS = physical component summary. MCS = mental component summary.
M. Ørbo et al. / Resuscitation 105 (2016) 92–99
97
Table 3 Frequency of impaired scores according to the respective cut-offs and reliable change indexes. Cognitive measures
T1 Number of patients (%) scoring ≤ −1.5 SD
T2 Number of patients (%) scoring ≤ −1.5 SD
RCI T1–T2 Decline Number (%) of patients
RCI T1–T2 No change Number (%) of patients
RCI T1–T2 Improvement Number (%) of patients
Visual memory composite Verbal memory composite Psychomotor composite Executive composite WASI IQ
7 (21) 6 (18) 1 (3) 4 (12) 4 (12)
3 (10) 5 (15) 0 3 (10) 4 (12)
0 1 (3) 0 0 0
28 (85) 30 (91) 30 (91) 29 (88) 33 (100)
5 (15) 2 (6) 3 (10) 4 (12) 0
Self-report measures
T1 Number of patients (%) scoring above/under cut-offa
T2 Number of patients (%) scoring above/under cut-offa
RCI T1–T2 Decline Number (%) of patients
RCI T1–T2 No change Number (%) of patients
RCI T1–T2 Improvement Number (%) of patients
HADS anxiety HADS depression SF-36 PCS SF-36 MCS
3 (10) 0 11 (33) 3 (10)
4 (12) 5 (15) 7 (21) 6 (18)
2 (6) 6 (18) 0 5 (15)
31 (94) 26 (79) 29 (88) 28 (85)
0 1 (3) 4 (12) 0
T1 = 3 months post-arrest examination. T2 = 12 months post-arrest examination. SD = standard deviation. RCI = reliable change index. Decline was defined as an RCI score of Z = −1.645 or below, and improvement as an RCI score of Z = 1.645 or above.29 An RCI ± 1.645 indicates that 90% of change scores will be in this range and that an RCI > 1.645 will be observed by chance in only 5% and an RCI < −1.645 in 5%, assuming a normal distribution. WASI = Wechsler Abbreviated Scale of Intelligence. HADS = Hospital Anxiety Depression Scale. SF-36 = Short Form 36. PCS = physical component summary. MCS = mental component summary. a Cut-off for HADS anxiety and depression subscales are eight points, separately24 and a T-score below 40 was used as cut-off for the component summary scores on SF-36.26
Correlations between cognitive performance at T1 and SF-36, HADS and work status at T2
Discussion The present data suggest that there was stability in cognitive performance between 3 and 12 months after OHCA. Only minor spontaneous cognitive recovery was indicated by our results, implying that longer-term cognitive prognosis can be estimated from assessments conducted rather early after survival from OHCA.3,8,10,12 In addition to the group analyses, we conducted analyses on the reliable statistical improvement in each participant’s cognitive scores. These results suggested that statistically reliable improvements were possible in at least some cognitive domains in some individuals.11 A third of the participants showed reliable improvement in at least one cognitive domain. The reliable change index (RCI) is considered statistically conservative when a cut-off of 1.65 is used to determine change.29 From a clinical point of view, the magnitude of improvement required for reliable change to be observed in the neuropsychological data would often be considered a clinically relevant improvement, potentially outweighing the amount of improvement expected from test-retest
Table 4, panel A shows that a better performance on verbal memory tests was associated with a better HRQL (both PCS and MCS) and having returned to work. A better performance on the executive composite score was correlated with a better mental HRQL. A better verbal memory score was correlated with returning to work. There were no significant correlations between cognitive composites and HADS scores. Table 4, panel B shows correlations between HADS, SF-36 and returning to work at T2. The correlations between HADS depression and SF-36 summary component scores were significant, showing that more depressive symptoms were associated with both lower physical and mental HRQL. Increased anxiety was correlated with reduced mental HRQL but not physical HRQL. There were no significant correlations between HADS anxiety or depression scale scores and the return-to-work variable. Active workers reported better physical HRQL.
Table 4 Correlations between cognitive composites, Hospital Anxiety Depression Scale (HADS) scores, SF-36 main component scores and returned to work. A
Visual memory composite T1
Verbal memory composite T1
Psychomotor composite T1
Executive composite T1
HADS anxiety T2 HADS Depression T2 SF-36 PCS T2 SF-36 MCS T2 Returned to work T2
−.02 −.14 .28 .03 .22
−.02 −.12 .43* .20 .37*
−.20 −.21 .38* .26 .22
−.25 −.32 .32 .38* .31
B HADS anxiety T2 HADS depression T2 SF-36 PCS T2 SF-36 MCS T2 Returned to work T2
HADS anxiety T2 – .77** −.35 −.82** −.16
HADS depression T2 **
.77 – −.57** −.84** −.31
SF-36 PCS T2
SF-36 MCS T2
−.35 −.57** – .49* .40*
−.82** −.84** .49* – .35
Panel A shows the correlations between the cognitive composites, HADS scores, SF-36 summary component scores and work status. Panel B shows the correlations between HADS scores, SF-36 component summary scores and work status. T1 = three months post-arrest assessment. T2 = twelve months post-arrest assessment. HADS = Hospital Anxiety and Depression Scale. SF-36 = Short-Form 36. PCS = physical component summary. MCS = mental component summary. * p < .05. ** p < .01. p-Values are uncorrected for multiple testing. N = 33.
98
M. Ørbo et al. / Resuscitation 105 (2016) 92–99
effects alone.29,30 Still, training effects are not controlled for by the RCIs used.29,30 The present study lacked power, which prohibits certain conclusions about the alterations in cognitive outcome, the modeling of variables associated with more or less improvement in cognitive results over time and the ability to compare subgroups of participants with different resuscitation and outcome characteristics. Furthermore, the present sample was biased toward OHCA survivors who had favorable prognostic characteristics.31 Hence, the frequency of survivors with impaired cognitive test performances and elevated psychological distress was lower than in other studies.1,3,6,14 Survivors who were awake at admission or who had only brief periods of coma after admission were not excluded from the present analyses. This is a subgroup of patients who are usually expected to recover to their premorbid neurologic and cognitive functioning,10,32 and this subgroup was excluded from some9,10 but not all8 of the previous prospective studies with a repeated cognitive assessment. The present data showed that a longer duration of coma was associated with worse cognitive outcome at both time points of assessment. The nine survivors who were awake at admission did not show statistically better cognitive performance or HRQL or less psychological distress. However, only one of the survivors awake at admission scored in the cognitively impaired range. Poorer cognitive functioning at 3 months was correlated with worse physical and mental HRQL and not having returned to work in the first year after OHCA. Similar associations between objective cognitive performance and HRQL have been found in other studies10,13 and have also been reported for subjective cognitive complaints.6,14 This shows the importance of cognitive functioning in HRQL after OHCA.6,10,13,14 The cognitive results were not, however, associated with depression or anxiety scores in the present study. The detrimental effects of anxiety and depression on HRQL were illustrated by the significant correlations observed between HADS and SF-36 scores in the present study. These results have been substantiated by larger and more controlled studies.14,33 Elevated psychological distress has been shown to negatively influence both physical health and HRQL in cardiac populations.14,33 Depressive symptoms and mental HRQL worsened over time. This finding contrasts with the results of Larsson et al., who suggested improvements in HRQL with time and stable reports of psychological distress, at least in the first 6 months after cardiac arrest.34 However, the present results underscore the argument made by Raina et al.,35 that combining multiple outcome measures is necessary to obtain complete information about different aspects of functional recovery.35 Due to the methodological limitations in this observational study, the results should be viewed as descriptive and hypothesis generating rather than conclusive. Of particular concern is the loss of participants due to various reasons and lack of a control group without cardiac arrest but with similar cardiovascular burden.1,2,5,10 The strengths are the prospective inclusion, that all participants had OHCA due to a presumed cardiac origin, the two time points of measurement up to one year after OHCA and the assessment of cognitive functioning with performance-based, psychometrically validated tests with the ability to capture even subtle cognitive impairments.3 Currently, there are limited data that can inform clinical practice and research regarding the optimal time point of when to perform follow-up assessment(s) and which exact outcome measures provide the most complete picture of functional outcomes after OHCA.35,36 The present data support the supposition that both cognitive performance and psychological distress are important outcome variables, as both influence longer-term HRQL.6,7,10,13,14,33 Existing studies suggest that results from cognitive assessments performed early after resuscitation can approximate longer-term
outcomes.3,10 While systematic, early screening that includes assessment of cognitive performance and psychological distress has been recommended in recent clinical guidelines for postresuscitation care, standardized clinical follow-up routines that incorporate these concepts are not implemented in most places.7,37 Conflict of interest statement The authors declare that there is no conflict of interest. Acknowledgements This work was supported by the Norwegian Extra-Foundation for Health and Rehabilitation through EXTRA funds; the Northern Norway Regional Health Authority; and the University Hospital of North Norway and The Arctic University of Norway, UiT, Tromsø. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.resuscitation. 2016.05.017. References 1. Lilja G, Nielsen N, Friberg H, et al. Cognitive function in survivors of out-ofhospital cardiac arrest after target temperature management at 33 ◦ C versus 36 ◦ C. Circulation 2015;131:1340–9. 2. Alexander MP, Lafleche G, Schnyer D, Lim C, Verfaellie M. Cognitive and functional outcome after out of hospital cardiac arrest. J Int Neuropsychol Soc 2011;17:364–8. 3. Moulaert VR, Verbunt JA, van Heugten CM, Wade DT. Cognitive impairments in survivors of out-of-hospital cardiac arrest: a systematic review. Resuscitation 2009;80:297–305. 4. Orbo M, Aslaksen PM, Larsby K, et al. Determinants of cognitive outcome in survivors of out-of-hospital cardiac arrest. Resuscitation 2014;85:1462–8. 5. Cronberg T, Lilja G. Cognitive decline after cardiac arrest – it is more to the picture than hypoxic brain injury. Resuscitation 2015;91:A3–4. 6. Moulaert VR, Wachelder EM, Verbunt JA, Wade DT, van Heugten CM. Determinants of quality of life in survivors of cardiac arrest. J Rehabil Med 2010;42:553–8. 7. Nolan JP, Soar J, Cariou A, et al. European Resuscitation Council and European Society of Intensive Care Medicine Guidelines for Post-resuscitation Care 2015: Section 5 of the European Resuscitation Council Guidelines for Resuscitation 2015. Resuscitation 2015;95:202–22. 8. Sauve MJ, Walker JA, Massa SM, Winkle RA, Scheinman MM. Patterns of cognitive recovery in sudden cardiac arrest survivors: the pilot study. Heart Lung 1996;25:172–81. 9. Roine RO, Kajaste S, Kaste M. Neuropsychological sequelae of cardiac arrest. JAMA 1993;269:237–42. 10. Lim C, Verfaellie M, Schnyer D, Lafleche G, Alexander MP. Recovery, long-term cognitive outcome and quality of life following out-of-hospital cardiac arrest. J Rehabil Med 2014;46:691–7. 11. Lim C, Alexander MP, LaFleche G, Schnyer DM, Verfaellie M. The neurological and cognitive sequelae of cardiac arrest. Neurology 2004;63:1774–8. 12. Drysdale EE, Grubb NR, Fox KA, O’Carroll RE. Chronicity of memory impairment in long-term out-of-hospital cardiac arrest survivors. Resuscitation 2000;47:27–32. 13. Orbo M, Aslaksen PM, Larsby K, et al. Relevance of cognition to health-related quality of life in good-outcome survivors of out-of-hospital cardiac arrest. J Rehabil Med 2015;47:860–6. 14. Lilja G, Nilsson G, Nielsen N, et al. Anxiety and depression among out-of-hospital cardiac arrest survivors. Resuscitation 2015;97:68–75. 15. Zigmond AS, Snaith RP. The hospital anxiety and depression scale. Acta Psychiatr Scand 1983;67:361–70. 16. Ware Jr JE, Sherbourne CD. The MOS 36-item short-form health survey (SF-36). I. Conceptual framework and item selection. Med Care 1992;30:473–83. 17. Delis DKE, Kramer J, Ober BT. California verbal learning test. 2nd ed. San Antonio, TX: The Psychological Corporation; 2000. 18. Delis DKE, Kramer J. Delis-Kaplan executive function systems. San Antonio, TX: The Psychological Corporation; 2001. 19. Heaton R, Miller S, Taylor M, Grant I. Norms for an expanded Halstead-Reitan battery: demographically adjusted neuropsychological norms for African American and Caucasian adults. Lutz: Psychol Assess Resour 2004. 20. Meyers J, Meyers KR. Rey complex figure test and recognition trial; 2003. 21. Wechsler D. Wechsler Abbreviated Scale of intelligence. San Antonio: The Psychological Corporation; 1999.
M. Ørbo et al. / Resuscitation 105 (2016) 92–99 22. Woods SP, Delis DC, Scott JC, Kramer JH, Holdnack JA. The California Verbal Learning Test – second edition: test–retest reliability, practice effects, and reliable change indices for the standard and alternate forms. Arch Clin Neuropsychol 2006;21:413–20. 23. Becker LB, Aufderheide TP, Geocadin RG, et al. Primary outcomes for resuscitation science studies: a consensus statement from the American Heart Association. Circulation 2011;124:2158–77. 24. Herrmann C. International experiences with the Hospital Anxiety and Depression Scale – a review of validation data and clinical results. J Psychosom Res 1997;42:17–41. 25. Loge JH, Kaasa S. Short form 36 (SF-36) health survey: normative data from the general Norwegian population. Scand J Soc Med 1998;26:250–8. 26. Ferguson RJ, Robinson AB, Splaine M. Use of the reliable change index to evaluate clinical significance in SF-36 outcomes. Qual Life Res 2002;11:509–16. 27. Ferguson CJ. An effect size primer: a guide for clinicians and researchers. Prof Psychol-Res Pr 2009;40:532–8. 28. Jacobson NS, Truax P. Clinical significance: a statistical approach to defining meaningful change in psychotherapy research. J Consult Clin Psychol 1991;59:12–9. 29. Duff K. Evidence-based indicators of neuropsychological change in the individual patient: relevant concepts and methods. Arch Clin Neuropsychol 2012;27:248–61.
99
30. Heilbronner RL, Sweet JJ, Attix DK, Krull KR, Henry GK, Hart RP. Official position of the American Academy of Clinical Neuropsychology on serial neuropsychological assessments: the utility and challenges of repeat test administrations in clinical and forensic contexts. Clin Neuropsychol 2010;24:1267–78. 31. Goossens PH, Moulaert VR. Cognitive impairments after cardiac arrest: implications for clinical daily practice. Resuscitation 2014;85:A3–4. 32. Festa JLR. Neurovascular neuropsychology. NY, USA: Springer; 2009. 33. Moulaert VR, van Heugten CM, Winkens B, et al. Early neurologically-focused follow-up after cardiac arrest improves quality of life at one year: a randomised controlled trial. Int J Cardiol 2015;193:8–16. 34. Larsson IM, Wallin E, Rubertsson S, Kristofferzon ML. Health-related quality of life improves during the first six months after cardiac arrest and hypothermia treatment. Resuscitation 2014;85:215–20. 35. Raina KD, Rittenberger JC, Holm MB, Callaway CW. Functional outcomes: one year after a cardiac arrest. Biomed Res Int 2015;2015:283608. 36. Smith K, Bernard S. Quality of life after cardiac arrest: how and when to assess outcomes after hospital discharge? Resuscitation 2014;85:1127. 37. Israelsson J, Lilja G, Bremer A, Stevenson-Agren J, Arestedt K. Post cardiac arrest care and follow-up in Sweden – a national web-survey. BMC Nurs 2016; 15:1.
Contents lists available at ScienceDirect
Resuscitation journal homepage: www.elsevier.com/locate/resuscitation
Clinical paper
Alterations in cognitive outcome between 3 and 12 months in survivors of out-of-hospital cardiac arrest Marte Ørbo a,∗ , Per M. Aslaksen b,c , Kristina Larsby d , Christoph Schäfer a , Pål M. Tande d , Audny Anke a,f a
Department of Rehabilitation, Division of Neurosciences, Orthopedics and Rehabilitation Services, University Hospital of North Norway, Norway Department of Psychology, Faculty of Health Sciences, University of Tromsø, Norway c Department of Child and Adolescent Psychiatry, Division of Child and Adolescent Health, The Regional Unit for Eating Disorders, University Hospital of North Norway, Norway d Department of Cardiology, Division of Cardiothoracic and Respiratory Medicine, University Hospital of North Norway, Norway f Department of Clinical Medicine, Faculty of Health Sciences, University of Tromsø, Norway b
a r t i c l e
i n f o
Article history: Received 25 September 2015 Received in revised form 17 April 2016 Accepted 15 May 2016 Keywords: Cardiac arrest Cognitive Rehabilitation Recovery Neuropsychology
a b s t r a c t Objectives: To prospectively investigate cognitive recovery from 3 to 12 months after resuscitation from out-of-hospital cardiac arrest (OHCA) and the associations between cognitive performance at 3 months and health-related quality of life (HRQL), psychological distress and work status after 12 months. Methods: At both assessments, neuropsychological tests were used to measure aspects of general mental ability, verbal and visual memory, psychomotor speed and executive function. The Short Form-36 (SF-36) was used to measure mental and physical HRQL, and the Hospital Anxiety and Depression Scale (HADS) to assess psychological distress. Results: 33 survivors completed both exams (31 males, mean age 58.6 years, SD = 13). The OHCAs were witnessed and due to cardiac origins. Nine patients were awake at admission to the hospital. Longer coma duration was associated with poorer cognitive results. Memory impairments were the most common symptom. The mean changes and effect sizes indicated minor improvements in cognitive performance from 3 to 12 months (Hedges g ≤ .26). Reliable change indices for an individual’s results further confirmed the stability of the group statistics. The HADS scores showed increased depressive symptoms, and mental HRQL was reduced from 3 to 12 months. Higher reports of psychological distress were related to worse HRQL. Work participation increased. Better cognitive results at 3 months were correlated with better HRQL and return to work at 12 months. Conclusions: The current data describe stability in results from 3 to 12 months. A worse cognitive performance at 3 months and higher reports of psychological distress were associated with lower HRQL. © 2016 Elsevier Ireland Ltd. All rights reserved.
Introduction Cognitive dysfunction is estimated to occur in 30–50% of survivors of out-of-hospital cardiac arrest (OHCA) when measured with objective, performance-based neuropsychological tests in the first year after resuscitation.1–3 Most frequently identified are problems with memory functions.1–4 In addition to the hypoxic-ischemia caused by the circulatory arrest, other risk factors for cognitive decline exist in this patient group and include
∗ Corresponding author at: Department of Rehabilitation, Division of Neurosciences, Orthopedics and Rehabilitation Services, University Hospital of North Norway, 9038 Tromsø, Norway. E-mail address: [email protected] (M. Ørbo). http://dx.doi.org/10.1016/j.resuscitation.2016.05.017 0300-9572/© 2016 Elsevier Ireland Ltd. All rights reserved.
pre-existing cardiovascular burden and general critical illness.1,5 Despite the existence of multiple risk factors for cognitive problems, cognitive decline is not regularly assessed and may go unrecognized in post-cardiac arrest care and follow-up.3,6,7 Previous studies of cognitive outcomes have most often relied on a single time point of measurement.3 Few previous prospective studies have assessed improvement in cognitive performance based on repeated neuropsychological testing of the same survivors over time.3,8–12 It is generally suggested that cognitive recovery plateaus in the first 3 months after resuscitation and that early cognitive assessments approximate longer-term outcomes.3 The aim of the present study was to assess improvement in cognitive functioning 3–12 months after resuscitation. Additionally, this study aimed to explore if worse cognitive functioning detected at 3 months anticipated worse functional prognosis in terms of poorer health related
M. Ørbo et al. / Resuscitation 105 (2016) 92–99
quality of life (HRQL), more psychological distress or not returning to work 1 year after an OHCA. Based on the results from previous studies, we expected little cognitive improvement beyond 3 months3,10 but believed that lower cognitive performance could be associated with lower HRQL after 1 year.6,10,13 We further expected that higher reported levels of psychological distress would negatively influence HRQL.6,14 Methods The study was approved by the Regional Committee for Medical Research Ethics in North-Norway, institutional protocol number 2009/1395. Written consent was obtained from all participants. Participants We aimed to prospectively include all survivors of nontraumatic, normothermic OHCA of presumed cardiac origin, aged 18–85 years, who were discharged from the cardiac ward at the University hospital of North Norway between August 2010 and September 2013 and who were able to undergo a neuropsychological assessment 3 months after resuscitation. The hospital is a tertiary care center with an advanced cardiac ward serving smaller hospitals in a regional manner. According to hospital records, 197 survivors were treated for OHCA of any cause in the study period. Patients who died before or during hospital admission were not registered. Of these OHCAs, 129 were of cardiac origin. Participants had to be fluent Norwegian speakers, be living independently prior to cardiac arrest and have no history of psychiatric or neurologic disease or significant ongoing somatic disease, substance abuse or alcoholism. Fig. 1 shows a flow-chart starting with the total number of OHCA patients discharged alive from the cardiac ward during the study period who had an OHCA due to a presumed cardiac origin; the reasons for exclusion and loss to follow-up at both time points for assessment are also displayed. Procedure Survivors were invited to the hospital for an individual, outpatient, face-to-face follow-up assessment at 3 months (T1) and at 12 months (T2) post-OHCA that included performance-based neuropsychological tests and 2 self-report questionnaires: the Hospital Anxiety and Depression Scale (HADS)15 and the Short Form (36) Health Survey (SF-36) version 1.2.16 Experienced neuropsychologists administered and scored all tests according to standardized procedures. Tests and questionnaires were administered in one session at each time point. Age, education history, living situation and working status prior to cardiac arrest and at both follow-up time points were recorded from the patient interview. Information about their medical history and treatment variables were recorded from participants’ medical journals at the hospital. Description of the outcome measures Neuropsychological tests Details of the individual neuropsychological tests17–22 are described in Supplementary Table 1. We selected tests that were representative of the cognitive functions described as being the most frequently impaired in OHCA survivors: declarative memory, executive function and psychomotor function.2,3,11,13,23 The study was performed in a clinical setting, and thus only tests that are commonly applied in clinical practice were used. The tests selected have been widely used internationally and have been standardized to the Norwegian setting, with psychometric properties such as construct validity and test-retest reliability that have empirically been shown to be satisfactory to excellent. Published normative data from large groups of healthy individuals were used to
93
score each test correcting for demographic variables.17–22 The standardized test results were transformed to Z-scores: Patient score − normative mean score Normative standard deviation of the test The tests were grouped into five cognitive domain composite scores based on construct validity. A mean Z-score for each domain was computed to both reduce the number of comparisons in the statistical analyses and to reduce the influence of possible idiosyncratic results that may have occurred by chance on single tests. A cut-off of ≤−1.5 SD from the normative mean was used to characterize abnormal performance on each cognitive composite score. Self-report questionnaires The HADS is a brief self-report questionnaire that consists of 14 questions; 7 questions assess anxiety symptoms and 7 assess symptoms of depression (score 0–3 for each question), each scale having a maximum score of 21.24 Scores at or above eight points on either the depression or anxiety scale are suggested to be clinically relevant.14,24 The questionnaire is frequently used in studies of cardiac arrest survivors and has been shown to have sound psychometric properties.14 HRQL was measured with the Medical Outcomes Study 36Item Short Form Health Survey (SF-36 version 1.2), a generic HRQL measure that consists of 2 summary scales, a physical component scale (PCS) and a mental component scale (MCS), which are divided into 8 subscales: physical functioning (PF), physical role (RP), bodily pain (BP), general health (GH), mental health (MH), emotional role (RE), social function (SF) and vitality (VT).16,25,26 The questionnaire asks patients about their perceptions of how their health status has interfered with their psychological, social and physical functioning in the previous 4 weeks. The participants’ data were compared with age- and gender-corrected normative data from the general Norwegian population.25 An online calculator for norm-based data scoring (http://www.sf36.org/nbscalc/index. shtlml) was applied for each survey, and the results were provided in normalized T-scores. When standardizing the scores according to the T-score distribution, the means were 50 and the standard deviations were 10 across all summary scales and subscales on the SF-36 in the comparator group. Thus, a direct comparison between the participants’ T-scores and the age- and gender-corrected normative data in the general Norwegian population could be viewed in a single graph. Fig. 2 shows the standardized SF-36 results from the 33 participants at both time points of assessment. T < 40 on any of the scales is suggested to represent HRQL below the normal range.26 Statistical analyses Statistical analyses were performed in IBM SPSS (version 22). Missing data at T1 were found in three participants on the Rey Complex Figure and in two participants on the Grooved Pegboard tests. At T2, missing data were found for one participant on the Grooved Pegboard test and in two participants on the WordFluency Test. In the HADS data, two participants had missing data. All missing data were replaced by the series mean. Before performing inferential statistics, all continuous variables were examined for normality by the Shapiro–Wilk test and by visual inspection of the Q–Q plots and box-plots of the distributions. Continuous variables were analyzed by parametric tests, as none of the continuous data were significantly different from a normal distribution. Statistical group comparisons were performed by one-way analysis of variance (ANOVA). To test whether the mean of the continuous data differed from normative data, one-sample t-tests were
94
M. Ørbo et al. / Resuscitation 105 (2016) 92–99
Discharged alive after OHCA due to cardiac cause August 2010 – September 2013 Excluded due to predefined criteria N = 129
Age above 85 years, n = 2 Severe anoxic brain injury, n = 8 Tourists/living abroad, n = 10 Non-fluent in Norwegian language, n = 1 Other severe somatic illness, n = 7 Psychiatric disorder / neurological disease / learning disabilities / drug/alcohol abuse, n = 18
Died before 3 months after OHCA n=4 Eligible for the study N = 79
Lost to follow-up Transferred to further treatment and rehabilitation/mild anoxic injury suspected, n = 6 Discharged to own home – no brain injury suspected and not asked to participate, n = 13
Refused to participate Reasons: Long travelling distance, n = 2 Included at 3 months
The examination felt unnecessary due to selfreported good functioning, n = 10
N = 45
Did not want to participate in research, n = 2 No show, n = 1
Participants lost from 3 to 12 months Reasons: Felt no need for a second evaluation due to self-reported good functioning, n = 3 Felt no use of a second evaluation, n = 1 Not asked to participate due to time restrictions in the data collection, n = 5. Logistics, n = 2 Somatic illness, n = 1 Included at 12 months N=33
Fig. 1. Overview of enrollment and exclusion. In addition to the predefined exclusion criteria shown in this figure, previous cardiac arrest and cardiac surgery during the year prior to OHCA were set as predefined reasons for exclusion. None of the patients included had experienced previous cardiac arrest nor had they received cardiac surgery in the year prior to OHCA.4
employed. Group comparisons with several dependent variables were performed by multivariate ANOVA. The analyses of changes from T1 to T2 in neuropsychological composite scores, HADS scales scores and SF-36 summary component scores were first performed with paired-sample t-tests. Then, the Holm-Bonferroni method was used to adjust the p-values
for multiple comparisons in the analyses assessing change from T1 to T2. In all other analyses, a p-value < .05 was considered significant. Hedges’ g was calculated to display the effect sizes of the mean change in continuous variables from 3 to 12 months post-OHCA. Effects were considered small when g was <.5, medium for g < .5–.8,
M. Ørbo et al. / Resuscitation 105 (2016) 92–99
95
Fig. 2. Comparisons between the subscale and summary component scores of the SF-36 at T1 and T2. The results from T1 are displayed first in each pair. Higher scores indicate better HRQL. SF-36 scores are shown in normed T-scores, with a mean = 50 and SD (standard deviation) = 10. Error bars denote SD. The solid line indicates the normative mean, and the two dotted lines indicate ±1 SD from the normative mean. PF = physical functioning. RP = role physical. BP = body pain. GH = general health. VT = vitality. SF = social functioning. RE = role emotional. MH = mental health. PCS = physical component summary. MCS = mental component summary. N = 33. * indicates significant difference (p < .05) tested by pairwise t-tests, uncorrected for multiple comparisons.
and large for g > .8.27 The change in frequency data (work status) was examined by the related samples McNemar’s change test. Reliable change indexes (RCI)28 were calculated for the neuropsychological domain composites, SF-36 summary components (PCS and MCS) and the HADS anxiety and depression scales according to the following formula:
T2 score − T1 score
2SDT12 (1 − correlation between T1 and T2)
An RCI-value indicating a change of ±1.645 SD was considered a statistically reliable change.29 Correlations between the neuropsychological composites and self-report measures were performed by two-tailed Pearson correlations. Results Sample characteristics The demographic, medical and resuscitation characteristics are displayed in Table 1. One-way ANOVA detected no group differences in duration of coma, age, years of education, ICD, hypothermia, time to ROSC or employment status, cognitive tests at T1 or self-report questionnaires at T1 (all Fs < 2.9, all p values > 0.1) between participants having neuropsychological data at both time points (N = 33) and in those who only had data from T1 (N = 12). The participants had variable coma durations, and nine were awake upon admission to the hospital. A multivariate ANOVA showed no significant differences in the cognitive composite scores, HADS scores or SF-36 MCS or PCS scores between participants who were awake at admission and those who were unconscious (all Fs < 3.95). Data for the participants who were awake at admission compared to those who were unconscious at admission are shown in Supplementary Table 2. Only one of the patients awake at admission scored in the cognitively impaired range. A longer coma duration was significantly associated with poorer performance in all cognitive domains at T1 and T2 (all p values <.01 adjusted for multiple comparisons, all Fs > 5.36) but not with verbal memory at T2.
Descriptive results at T1 and T2 Table 2 presents the mean, SD and range for the five cognitive composite scores, the HADS depression and anxiety scores and the SF-36 summary component scores (PCS and MCS) at both T1 and T2. Additionally, descriptive information of all SF-36 subscale scores are provided in Fig. 2. Table 3 shows the percentage of abnormal scores according to the defined cut-off points for the HADS scales, SF-36 summary component scores and cognitive composite scores. Work status is shown in Table 1. Assessment of changes in outcome variables from T1 to T2 Table 2 shows the results of the analyses of statistical changes in the cognitive composite, SF-36 summary component and HADS scores between T1 and T2 in terms of t-tests and effects sizes. For cognitive outcomes, the paired t-tests suggested significant improvement in two cognitive domains: visual memory (t = −2.46, p = .02) and executive function (t = −2.31, p = .03). The effect sizes were small (Hedges g ≤ .26). When lowering the significance level according to the Holm-Bonferroni adjustment for multiple comparisons, a significant improvement was shown in the visual memory domain only (p < .025). A significant decline was found from T1 to T2 in the MCS scale score on the SF-36, indicating a worsening of mental HRQL. When examining changes in the 8 different SF-36 subscales that constitute the summary components, no significant statistical change was revealed between T1 and T2 (all ps > .063) by two-tailed t-tests (Fig. 2). Table 2 further shows a significant increase in HADS depression but not anxiety scores from T1 to T2. The number of active workers significantly increased from T1 to T2 (McNemar = 4.9, df = 1, p = .021). Table 3 shows the RCI indexes for cognitive composite scores, the HADS scores and SF-36 summary component scores. The additional information on cognitive outcomes provided was that eight persons showed reliable improvement in one cognitive domain, and three persons showed improvement in two domains, suggesting some reliable improvement in 33% of the participants.
96
M. Ørbo et al. / Resuscitation 105 (2016) 92–99
Table 1 Demographical and medical characteristics of the sample. Demographics
Number (%)
Male Female Age Married/partner Years of education Active workers before OHCA Active workers after 3 months Active workers after 12 months Living in own home at 3 months Living in own home at 12 months
31 (94) 2 (6)
Mean (SD)
Min–Max
58.6 (13)
18–82
12 (4)
5–21
16 (12)
1–60
42.9 (57)
2 min–192 h
4.8 (4.1) 11.4 (6.2)
1–14 5–36
28 (85) 18 (55) 2 (6) 11 (33) 32 (97) 33 (100)
Medical history No cardiovascular history Hypertension Myocardial infarction Heart failure Ischemia Diabetes mellitus
19 (58) 5 (15) 7 (21) 1 (3) 1 (3) 2 (6)
OHCA and treatment characteristics First rhythm VF AF Asystole Time to ROSC – minutes Cause of arrest Infarction Rhythm Unknown Awake at admission Duration of coma in hours Hypothermia PCI CABG ICD Medication
31 (94) 1 (3) 1 (3)
27 (82) 4 (12) 2 (6) 9 (27) Yes 14 (42)a Yes 26 (79) Yes 2 (6) Yes 10 (30) 23 (69)
Time variables Length of hospital stay Intensive care unit Cardiac ward Rehabilitation Days from OHCA to T1 Days from OHCA to T2
110 (15) 371 (33)
79–122 312–419
Overview of participant’s demographic, resuscitation and medical variables (N = 33). All had witnessed OHCA with bystander cardiopulmonary resuscitation (CPR) initiated. VF = ventricular fibrillation. AF = atrial fibrillation. ROSC = return of spontaneous circulation. PCI = percutaneous coronary intervention. ICD = implanted cardioverter defibrillator. CABG = Coronary Artery Bypass Grafting. T1 = assessment after 3 months. T2 = assessment after 12 months. a Reasons for not receiving therapeutic hypothermia were (N = 19): awake at admission (9), immediate CABG after admission (1), renal complications (1), uncertain state of consciousness at admission (2), pneumonia (1), aspiration pneumonia (1), long transportation distance to hospital (2), sepsis (1), procedural complication, hypothermia terminated before target temperature (1). Table 2 Assessment of change in outcomes from T1 to T2 (N = 33). Mean T1 (SD)
Range T1
Mean T2 (SD)
Range T2
r (T1–T2)
t
p
Cognitive composite Visual memory composite Verbal memory composite Psychomotor composite Executive composite WASI IQa
Hedges g
CI (95%) for g
−.60 (1.51) −.49 (1.13) −.01 (.64) −.13 (1.15) 98 (15.43)
−3 to 2.3 −3.4 to 1.8 −1.9 to 1.1 −3 to 1.3 60–126
.39 (1.57) −.27 (1.28) .15 (.53) .12 (.87) 99 (15.33)
−3 to 3 −3.5 to 1.6 −1.4 to 1.3 −2.7 to 1.7 70–127
.84** .80** .72** .83** .88**
−2.46 −1.65 −1.96 −2.31 −.78
.02† .11 .06 .03 .44
.26 .18 .25 .25 .07
−.25 to .76 −.30 to .66 −.24 to .74 −.24 to .75 −3.55 to 3.68
Self-report measures HADS anxiety HADS depression SF-36 PCSb SF-36 MCSb
2.94 (2.92) 1.7 (1.89) 44.10 (8.55) 51.85 (9.57)
0–11 0–5 29.7–56.9 29–63
3.78 (3.8) 2.56 (3.24) 45.82 (8.04) 47.71 (11.82)
0–16 0–12 29.7–57 18.4–63.4
.62** .80** .92** .70**
−1.71 −2.5 −1.58 2.60
.09 .02† .12 .02†
.27 .35 .21 −.38
−.55 to 1.09 −.30 to .99 −1.80 to 2.21 −2.98 to 2.21
Paired samples t-tests and effects sizes of the change (Hedges g). Cognitive composites are displayed as Z-scores (mean = 0, SD = 1). Higher Z-scores indicate better performance. r = Pearson correlation. ** p < .001. † Significant change after correction for multiple comparisons by the Holm-Bonferroni procedure, adjusting for correlation between the measures and degrees of freedom. Adjustments for multiple comparisons were performed separately for cognitive (p < .025) and self-report measures (p < .035). Mean correlation between the cognitive composites was r = .56, and mean correlation between the self-report measures was r = .76. p values indicate difference in paired-sample t-tests. CI = Confidence interval. a WASI IQ is displayed as Wechsler scores, normative mean = 100, SD = 15. Higher scores indicate better performance. HADS = Hospital Anxiety Depression Scale, higher scores on both subscales indicates more symptoms. SF-36 = Short Form 36. b SF-36 is displayed in T-scores (normative mean = 50, SD = 10), higher scores indicates better HRQL. PCS = physical component summary. MCS = mental component summary.
M. Ørbo et al. / Resuscitation 105 (2016) 92–99
97
Table 3 Frequency of impaired scores according to the respective cut-offs and reliable change indexes. Cognitive measures
T1 Number of patients (%) scoring ≤ −1.5 SD
T2 Number of patients (%) scoring ≤ −1.5 SD
RCI T1–T2 Decline Number (%) of patients
RCI T1–T2 No change Number (%) of patients
RCI T1–T2 Improvement Number (%) of patients
Visual memory composite Verbal memory composite Psychomotor composite Executive composite WASI IQ
7 (21) 6 (18) 1 (3) 4 (12) 4 (12)
3 (10) 5 (15) 0 3 (10) 4 (12)
0 1 (3) 0 0 0
28 (85) 30 (91) 30 (91) 29 (88) 33 (100)
5 (15) 2 (6) 3 (10) 4 (12) 0
Self-report measures
T1 Number of patients (%) scoring above/under cut-offa
T2 Number of patients (%) scoring above/under cut-offa
RCI T1–T2 Decline Number (%) of patients
RCI T1–T2 No change Number (%) of patients
RCI T1–T2 Improvement Number (%) of patients
HADS anxiety HADS depression SF-36 PCS SF-36 MCS
3 (10) 0 11 (33) 3 (10)
4 (12) 5 (15) 7 (21) 6 (18)
2 (6) 6 (18) 0 5 (15)
31 (94) 26 (79) 29 (88) 28 (85)
0 1 (3) 4 (12) 0
T1 = 3 months post-arrest examination. T2 = 12 months post-arrest examination. SD = standard deviation. RCI = reliable change index. Decline was defined as an RCI score of Z = −1.645 or below, and improvement as an RCI score of Z = 1.645 or above.29 An RCI ± 1.645 indicates that 90% of change scores will be in this range and that an RCI > 1.645 will be observed by chance in only 5% and an RCI < −1.645 in 5%, assuming a normal distribution. WASI = Wechsler Abbreviated Scale of Intelligence. HADS = Hospital Anxiety Depression Scale. SF-36 = Short Form 36. PCS = physical component summary. MCS = mental component summary. a Cut-off for HADS anxiety and depression subscales are eight points, separately24 and a T-score below 40 was used as cut-off for the component summary scores on SF-36.26
Correlations between cognitive performance at T1 and SF-36, HADS and work status at T2
Discussion The present data suggest that there was stability in cognitive performance between 3 and 12 months after OHCA. Only minor spontaneous cognitive recovery was indicated by our results, implying that longer-term cognitive prognosis can be estimated from assessments conducted rather early after survival from OHCA.3,8,10,12 In addition to the group analyses, we conducted analyses on the reliable statistical improvement in each participant’s cognitive scores. These results suggested that statistically reliable improvements were possible in at least some cognitive domains in some individuals.11 A third of the participants showed reliable improvement in at least one cognitive domain. The reliable change index (RCI) is considered statistically conservative when a cut-off of 1.65 is used to determine change.29 From a clinical point of view, the magnitude of improvement required for reliable change to be observed in the neuropsychological data would often be considered a clinically relevant improvement, potentially outweighing the amount of improvement expected from test-retest
Table 4, panel A shows that a better performance on verbal memory tests was associated with a better HRQL (both PCS and MCS) and having returned to work. A better performance on the executive composite score was correlated with a better mental HRQL. A better verbal memory score was correlated with returning to work. There were no significant correlations between cognitive composites and HADS scores. Table 4, panel B shows correlations between HADS, SF-36 and returning to work at T2. The correlations between HADS depression and SF-36 summary component scores were significant, showing that more depressive symptoms were associated with both lower physical and mental HRQL. Increased anxiety was correlated with reduced mental HRQL but not physical HRQL. There were no significant correlations between HADS anxiety or depression scale scores and the return-to-work variable. Active workers reported better physical HRQL.
Table 4 Correlations between cognitive composites, Hospital Anxiety Depression Scale (HADS) scores, SF-36 main component scores and returned to work. A
Visual memory composite T1
Verbal memory composite T1
Psychomotor composite T1
Executive composite T1
HADS anxiety T2 HADS Depression T2 SF-36 PCS T2 SF-36 MCS T2 Returned to work T2
−.02 −.14 .28 .03 .22
−.02 −.12 .43* .20 .37*
−.20 −.21 .38* .26 .22
−.25 −.32 .32 .38* .31
B HADS anxiety T2 HADS depression T2 SF-36 PCS T2 SF-36 MCS T2 Returned to work T2
HADS anxiety T2 – .77** −.35 −.82** −.16
HADS depression T2 **
.77 – −.57** −.84** −.31
SF-36 PCS T2
SF-36 MCS T2
−.35 −.57** – .49* .40*
−.82** −.84** .49* – .35
Panel A shows the correlations between the cognitive composites, HADS scores, SF-36 summary component scores and work status. Panel B shows the correlations between HADS scores, SF-36 component summary scores and work status. T1 = three months post-arrest assessment. T2 = twelve months post-arrest assessment. HADS = Hospital Anxiety and Depression Scale. SF-36 = Short-Form 36. PCS = physical component summary. MCS = mental component summary. * p < .05. ** p < .01. p-Values are uncorrected for multiple testing. N = 33.
98
M. Ørbo et al. / Resuscitation 105 (2016) 92–99
effects alone.29,30 Still, training effects are not controlled for by the RCIs used.29,30 The present study lacked power, which prohibits certain conclusions about the alterations in cognitive outcome, the modeling of variables associated with more or less improvement in cognitive results over time and the ability to compare subgroups of participants with different resuscitation and outcome characteristics. Furthermore, the present sample was biased toward OHCA survivors who had favorable prognostic characteristics.31 Hence, the frequency of survivors with impaired cognitive test performances and elevated psychological distress was lower than in other studies.1,3,6,14 Survivors who were awake at admission or who had only brief periods of coma after admission were not excluded from the present analyses. This is a subgroup of patients who are usually expected to recover to their premorbid neurologic and cognitive functioning,10,32 and this subgroup was excluded from some9,10 but not all8 of the previous prospective studies with a repeated cognitive assessment. The present data showed that a longer duration of coma was associated with worse cognitive outcome at both time points of assessment. The nine survivors who were awake at admission did not show statistically better cognitive performance or HRQL or less psychological distress. However, only one of the survivors awake at admission scored in the cognitively impaired range. Poorer cognitive functioning at 3 months was correlated with worse physical and mental HRQL and not having returned to work in the first year after OHCA. Similar associations between objective cognitive performance and HRQL have been found in other studies10,13 and have also been reported for subjective cognitive complaints.6,14 This shows the importance of cognitive functioning in HRQL after OHCA.6,10,13,14 The cognitive results were not, however, associated with depression or anxiety scores in the present study. The detrimental effects of anxiety and depression on HRQL were illustrated by the significant correlations observed between HADS and SF-36 scores in the present study. These results have been substantiated by larger and more controlled studies.14,33 Elevated psychological distress has been shown to negatively influence both physical health and HRQL in cardiac populations.14,33 Depressive symptoms and mental HRQL worsened over time. This finding contrasts with the results of Larsson et al., who suggested improvements in HRQL with time and stable reports of psychological distress, at least in the first 6 months after cardiac arrest.34 However, the present results underscore the argument made by Raina et al.,35 that combining multiple outcome measures is necessary to obtain complete information about different aspects of functional recovery.35 Due to the methodological limitations in this observational study, the results should be viewed as descriptive and hypothesis generating rather than conclusive. Of particular concern is the loss of participants due to various reasons and lack of a control group without cardiac arrest but with similar cardiovascular burden.1,2,5,10 The strengths are the prospective inclusion, that all participants had OHCA due to a presumed cardiac origin, the two time points of measurement up to one year after OHCA and the assessment of cognitive functioning with performance-based, psychometrically validated tests with the ability to capture even subtle cognitive impairments.3 Currently, there are limited data that can inform clinical practice and research regarding the optimal time point of when to perform follow-up assessment(s) and which exact outcome measures provide the most complete picture of functional outcomes after OHCA.35,36 The present data support the supposition that both cognitive performance and psychological distress are important outcome variables, as both influence longer-term HRQL.6,7,10,13,14,33 Existing studies suggest that results from cognitive assessments performed early after resuscitation can approximate longer-term
outcomes.3,10 While systematic, early screening that includes assessment of cognitive performance and psychological distress has been recommended in recent clinical guidelines for postresuscitation care, standardized clinical follow-up routines that incorporate these concepts are not implemented in most places.7,37 Conflict of interest statement The authors declare that there is no conflict of interest. Acknowledgements This work was supported by the Norwegian Extra-Foundation for Health and Rehabilitation through EXTRA funds; the Northern Norway Regional Health Authority; and the University Hospital of North Norway and The Arctic University of Norway, UiT, Tromsø. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.resuscitation. 2016.05.017. References 1. Lilja G, Nielsen N, Friberg H, et al. Cognitive function in survivors of out-ofhospital cardiac arrest after target temperature management at 33 ◦ C versus 36 ◦ C. Circulation 2015;131:1340–9. 2. Alexander MP, Lafleche G, Schnyer D, Lim C, Verfaellie M. Cognitive and functional outcome after out of hospital cardiac arrest. J Int Neuropsychol Soc 2011;17:364–8. 3. Moulaert VR, Verbunt JA, van Heugten CM, Wade DT. Cognitive impairments in survivors of out-of-hospital cardiac arrest: a systematic review. Resuscitation 2009;80:297–305. 4. Orbo M, Aslaksen PM, Larsby K, et al. Determinants of cognitive outcome in survivors of out-of-hospital cardiac arrest. Resuscitation 2014;85:1462–8. 5. Cronberg T, Lilja G. Cognitive decline after cardiac arrest – it is more to the picture than hypoxic brain injury. Resuscitation 2015;91:A3–4. 6. Moulaert VR, Wachelder EM, Verbunt JA, Wade DT, van Heugten CM. Determinants of quality of life in survivors of cardiac arrest. J Rehabil Med 2010;42:553–8. 7. Nolan JP, Soar J, Cariou A, et al. European Resuscitation Council and European Society of Intensive Care Medicine Guidelines for Post-resuscitation Care 2015: Section 5 of the European Resuscitation Council Guidelines for Resuscitation 2015. Resuscitation 2015;95:202–22. 8. Sauve MJ, Walker JA, Massa SM, Winkle RA, Scheinman MM. Patterns of cognitive recovery in sudden cardiac arrest survivors: the pilot study. Heart Lung 1996;25:172–81. 9. Roine RO, Kajaste S, Kaste M. Neuropsychological sequelae of cardiac arrest. JAMA 1993;269:237–42. 10. Lim C, Verfaellie M, Schnyer D, Lafleche G, Alexander MP. Recovery, long-term cognitive outcome and quality of life following out-of-hospital cardiac arrest. J Rehabil Med 2014;46:691–7. 11. Lim C, Alexander MP, LaFleche G, Schnyer DM, Verfaellie M. The neurological and cognitive sequelae of cardiac arrest. Neurology 2004;63:1774–8. 12. Drysdale EE, Grubb NR, Fox KA, O’Carroll RE. Chronicity of memory impairment in long-term out-of-hospital cardiac arrest survivors. Resuscitation 2000;47:27–32. 13. Orbo M, Aslaksen PM, Larsby K, et al. Relevance of cognition to health-related quality of life in good-outcome survivors of out-of-hospital cardiac arrest. J Rehabil Med 2015;47:860–6. 14. Lilja G, Nilsson G, Nielsen N, et al. Anxiety and depression among out-of-hospital cardiac arrest survivors. Resuscitation 2015;97:68–75. 15. Zigmond AS, Snaith RP. The hospital anxiety and depression scale. Acta Psychiatr Scand 1983;67:361–70. 16. Ware Jr JE, Sherbourne CD. The MOS 36-item short-form health survey (SF-36). I. Conceptual framework and item selection. Med Care 1992;30:473–83. 17. Delis DKE, Kramer J, Ober BT. California verbal learning test. 2nd ed. San Antonio, TX: The Psychological Corporation; 2000. 18. Delis DKE, Kramer J. Delis-Kaplan executive function systems. San Antonio, TX: The Psychological Corporation; 2001. 19. Heaton R, Miller S, Taylor M, Grant I. Norms for an expanded Halstead-Reitan battery: demographically adjusted neuropsychological norms for African American and Caucasian adults. Lutz: Psychol Assess Resour 2004. 20. Meyers J, Meyers KR. Rey complex figure test and recognition trial; 2003. 21. Wechsler D. Wechsler Abbreviated Scale of intelligence. San Antonio: The Psychological Corporation; 1999.
M. Ørbo et al. / Resuscitation 105 (2016) 92–99 22. Woods SP, Delis DC, Scott JC, Kramer JH, Holdnack JA. The California Verbal Learning Test – second edition: test–retest reliability, practice effects, and reliable change indices for the standard and alternate forms. Arch Clin Neuropsychol 2006;21:413–20. 23. Becker LB, Aufderheide TP, Geocadin RG, et al. Primary outcomes for resuscitation science studies: a consensus statement from the American Heart Association. Circulation 2011;124:2158–77. 24. Herrmann C. International experiences with the Hospital Anxiety and Depression Scale – a review of validation data and clinical results. J Psychosom Res 1997;42:17–41. 25. Loge JH, Kaasa S. Short form 36 (SF-36) health survey: normative data from the general Norwegian population. Scand J Soc Med 1998;26:250–8. 26. Ferguson RJ, Robinson AB, Splaine M. Use of the reliable change index to evaluate clinical significance in SF-36 outcomes. Qual Life Res 2002;11:509–16. 27. Ferguson CJ. An effect size primer: a guide for clinicians and researchers. Prof Psychol-Res Pr 2009;40:532–8. 28. Jacobson NS, Truax P. Clinical significance: a statistical approach to defining meaningful change in psychotherapy research. J Consult Clin Psychol 1991;59:12–9. 29. Duff K. Evidence-based indicators of neuropsychological change in the individual patient: relevant concepts and methods. Arch Clin Neuropsychol 2012;27:248–61.
99
30. Heilbronner RL, Sweet JJ, Attix DK, Krull KR, Henry GK, Hart RP. Official position of the American Academy of Clinical Neuropsychology on serial neuropsychological assessments: the utility and challenges of repeat test administrations in clinical and forensic contexts. Clin Neuropsychol 2010;24:1267–78. 31. Goossens PH, Moulaert VR. Cognitive impairments after cardiac arrest: implications for clinical daily practice. Resuscitation 2014;85:A3–4. 32. Festa JLR. Neurovascular neuropsychology. NY, USA: Springer; 2009. 33. Moulaert VR, van Heugten CM, Winkens B, et al. Early neurologically-focused follow-up after cardiac arrest improves quality of life at one year: a randomised controlled trial. Int J Cardiol 2015;193:8–16. 34. Larsson IM, Wallin E, Rubertsson S, Kristofferzon ML. Health-related quality of life improves during the first six months after cardiac arrest and hypothermia treatment. Resuscitation 2014;85:215–20. 35. Raina KD, Rittenberger JC, Holm MB, Callaway CW. Functional outcomes: one year after a cardiac arrest. Biomed Res Int 2015;2015:283608. 36. Smith K, Bernard S. Quality of life after cardiac arrest: how and when to assess outcomes after hospital discharge? Resuscitation 2014;85:1127. 37. Israelsson J, Lilja G, Bremer A, Stevenson-Agren J, Arestedt K. Post cardiac arrest care and follow-up in Sweden – a national web-survey. BMC Nurs 2016; 15:1.