0
Original Research |

Self-reported Depressive Symptoms and Memory Complaints in Survivors Five Years After ARDS201CDepression and Memory Complaints After ARDS201D FREE TO VIEW

Neill K. J. Adhikari, MDCM; Catherine M. Tansey, PhD; Mary Pat McAndrews, PhD; Andrea Matté, BSc; Ruxandra Pinto, PhD; Angela M. Cheung, MD, PhD; Natalia Diaz-Granados, MSc; Margaret S. Herridge, MD, MPH
Author and Funding Information

From the Interdepartmental Division of Critical Care (Drs Adhikari and Herridge), University of Toronto, Toronto; Department of Critical Care Medicine (Drs Adhikari and Pinto), Sunnybrook Health Sciences Centre, Toronto; Department of Medicine (Drs Tansey, Cheung, and Herridge and Ms Matté), Krembil Neuroscience Centre (Dr McAndrews), and Women’s Health Program (Dr Cheung and Ms Diaz-Granados), University Health Network, Toronto; Department of Medicine (Drs Cheung and Herridge) and Department of Health Policy, Management and Evaluation and the Dalla Lana School of Public Health (Dr Cheung), University of Toronto, Toronto; and Department of Clinical Epidemiology and Biostatistics (Ms Diaz-Granados), McMaster University, Hamilton, ON, Canada.

Correspondence to: Neill K. J. Adhikari, MDCM, Department of Critical Care Medicine, Room D1.08, Sunnybrook Health Sciences Centre, 2075 Bayview Ave, Toronto, ON, M4N 3M5, Canada; e-mail: neill.adhikari@utoronto.ca

Data are presented as median (first-third quartile) or No. (%). The 48 responders returned either a BDI-II or MAC-S questionnaire. APACHE = Acute Physiology and Chronic Health Evaluation; BDI-II = Beck Depression Inventory II; LIS = Lung Injury Score; MAC-S = Memory Assessment Clinics Self-Rating Scale; MODS = Multiple Organ Dysfunction Score.

a

n = 47.

b

The LIS included the sum of the chest radiography, hypoxemia, and positive end-expiratory pressure scores, while excluding static compliance.

c

n = 45.

d

n = 9.

Data are presented as median (first-third quartile) or No. (%). Percentages may not sum to 100% because of rounding. The two n values in the first column refer to the earlier questionnaires (completed at a median of 22 months [range, 6-48 months] following ICU discharge for BDI-II and MAC-S10 and 2 y for SF-36) and the 5-y questionnaires, respectively. Of 64 survivors evaluated at 5 y, 48 patients answered either the BDI-II or the MAC-S; data for the earlier time point includes only patients who also completed the 5-y questionnaires. Comparisons between 5-y and earlier scores were made using Wilcoxon signed rank test (continuous variables) and exact McNemar test (categorized scores) and only included patients who contributed data at both time points. MCS = mental component summary; MH = mental health; RE = role emotional; SF-36 = Medical Outcomes Study 36-Item Short Form. See Table 1 legend for expansion of other abbreviations.

a

At 5 y (n = 48), two patients did not answer BDI-II, and three had missing items; at the earlier administration, two did not answer, and three had missing items.

b

Depression categories are from the BDI-II scale. The P value refers to the comparison of minimal to mild vs moderate to severe categories at 5 y vs the earlier assessment.

c

At 5 years (n = 48), one patient did not answer, and one had a missing item; at the earlier administration, one did not answer, and five had missing items.

d

Proportion of sample < 2, < 1.5, or < 1 SD below age-adjusted US sample mean.14 We used patients’ ages at questionnaire completion.

e

At 5 y (n = 48), one patient did not answer, and nine had missing items; at the earlier assessment, one did not answer, and five had missing items.

f

At 5 y (n = 48), one patient did not answer, and three, one, and four patients had missing items for RE, MH, and MCS, respectively. At the earlier assessment, eight patients did not answer, and one and two patients had missing items for RE and MCS, respectively.

g

Proportion of sample < 2, < 1.5, or < 1 SD below age- and sex-matched Canadian sample mean.21 We used patients’ ages at questionnaire completion.

See Table 1 and 2 legends for expansion of abbreviations.

Predictor variables refer to the index ICU admission (except for age, which is at the time of questionnaire administration). Unadjusted analyses included questionnaires with no missing items (n = 43 except for APACHE II [n = 42]). Separately for each predictor, we also included the earlier BDI-II score and the number of months between BDI-II administrations (n = 38 except for APACHE II [n = 37]) in an adjusted analysis. All BDI-II scores were log-transformed. We added 0.5 to 0 scores before taking the logarithm (n = 4 at 5 y in unadjusted analyses; n = 3 at 5 y and n = 3 for the earlier questionnaire in adjusted analyses). Positive (negative) β coefficients imply that the predictor is associated with higher (lower) log-transformed BDI-II scores. In adjusted analyses, β values for the earlier BDI-II score ranged from 0.39 to 0.46 and were highly significant (P = .003-.015); β values for the time between administrations were small and not statistically significant. In sensitivity analyses calculating adjusted scores for questionnaires with missing items, results are similar. See Table 1 and 2 legends for expansion of abbreviations.

a

The change in MODS over time during ICU admission is expressed as the slope of the score.

b

The change in LIS over time during ICU admission is expressed as the slope of the score.

c

The logarithm of this variable was used because the untransformed variable had a skewed distribution.

Predictor variables refer to the index ICU admission (except for age, which is at the time of questionnaire administration). Unadjusted analyses include questionnaires with complete data (n = 46 for ability subscale except for APACHE II [n = 45]; n = 38 for frequency of occurrence subscale). Separately for each predictor, we also included the earlier MAC-S score and the number of months between MAC-S administrations in an adjusted analysis. For the ability subscale, adjusted analyses (n = 39 for all except APACHE II [n = 38]) excluded one case with high residuals. For the frequency of occurrence subscale, adjusted analyses (n = 30) excluded three cases with high residuals. Positive (negative) β coefficients imply that the predictor is associated with higher (lower) MAC-S scores. In adjusted analyses, β values for the earlier MAC-S score ranged from 0.84 to 0.94 and were highly significant (all P < .0001); β values for the time between administrations were small and not statistically significant. See Table 1 and 2 legends for expansion of abbreviations.

a

In a sensitivity analysis calculating adjusted scores for questionnaires with missing items, results were similar, but female sex was no longer significant in the adjusted analysis.

b

The change in MODS over time during ICU admission is expressed as the slope of the score.

c

The change in LIS over time during ICU admission is expressed as the slope of the score.

d

The logarithm of this variable was used because the untransformed variable had a skewed distribution.

e

In a sensitivity analysis including the three cases with high residuals, the adjusted analyses showed a significant effect of female sex (β, −11.60; SE, 4.60; P = .02). However, this effect was no longer significant if questionnaires with missing items were included by calculating adjusted scores.

Funding/Support: This work was performed at the University of Toronto and was supported by Physicians’ Services Incorporated, Ontario Thoracic Society, and Canadian Intensive Care Foundation. Dr Cheung is supported by a Canadian Institutes of Health Research Senior Investigator Award and the Lillian Love Chair in Women’s Health at the University of Toronto and University Health Network.

Reproduction of this article is prohibited without written permission from the American College of Chest Physicians (http://www.chestpubs.org/site/misc/reprints.xhtml).


Funding/Support: This work was performed at the University of Toronto and was supported by Physicians’ Services Incorporated, Ontario Thoracic Society, and Canadian Intensive Care Foundation. Dr Cheung is supported by a Canadian Institutes of Health Research Senior Investigator Award and the Lillian Love Chair in Women’s Health at the University of Toronto and University Health Network.

Funding/Support: This work was performed at the University of Toronto and was supported by Physicians’ Services Incorporated, Ontario Thoracic Society, and Canadian Intensive Care Foundation. Dr Cheung is supported by a Canadian Institutes of Health Research Senior Investigator Award and the Lillian Love Chair in Women’s Health at the University of Toronto and University Health Network.

Reproduction of this article is prohibited without written permission from the American College of Chest Physicians (http://www.chestpubs.org/site/misc/reprints.xhtml).

Reproduction of this article is prohibited without written permission from the American College of Chest Physicians (http://www.chestpubs.org/site/misc/reprints.xhtml).


Chest. 2011;140(6):1484-1493. doi:10.1378/chest.11-1667
Text Size: A A A
Published online

Background:  Survivors of ARDS report depressive symptoms and memory complaints, the prevalence of which after 5 years is unknown.

Methods:  We administered instruments assessing symptoms of depression (Beck Depression Inventory II [BDI-II]) and memory complaints (Memory Assessment Clinics Self-Rating Scale [MAC-S]) to 64 survivors of ARDS from four university-affiliated ICUs 5 years after ICU discharge. We compared BDI-II scores to quality of life (Medical Outcomes Study 36-Item Short Form [SF-36]) mental health domains (role emotional, mental health, mental component summary), compared BDI-II and MAC-S scores to earlier scores (median, 22 months postdischarge), and examined return to work.

Results:  Forty-three (67.2%), 46 (71.9%), and 38 (59.4%) patients fully completed the BDI-II, MAC-S ability subscale, and MAC-S frequency of occurrence subscale, respectively. Responders were young (median, 48 years; first-third quartile [Q1-Q3], 39-61 years) with high illness severity. The median BDI-II score was 10 (Q1-Q3, 3-18); eight of 43 (18.6%) had moderate to severe depressive symptoms compared with 14 of 43 (32.6%) earlier (P = .15, n = 38 with paired data). Median MAC-S ability and MAC-S frequency scores were 81 (Q1-Q3, 57-92) and 91.5 (Q1-Q3, 76-105), respectively, similar to earlier scores (P = .67 and P = .64, respectively); 0% to 4.3% scored > 2 SDs below population norms. Higher BDI-II score was predicted by higher earlier BDI-II score, slower recovery of organ function, and longer duration of mechanical ventilation and ICU stay. Higher MAC-S score was predicted by higher earlier MAC-S score. SF-36 mental health domain scores were very stable (P = .57-.83). BDI-II and SF-36 mental health domains were negatively correlated (Spearman coefficient, −0.50 to −0.82). Most patients returned to work regardless of depressive symptoms (minimal to mild, 31 of 35 [88.6%]; moderate to severe, five of eight [62.5%]; P = .12).

Conclusions:  Compared with ∼ 2 years postdischarge from the ICU, depressive symptoms and memory complaints were similar at 5 years. Mental health domains of the SF-36 may not be sensitive to small changes in mood symptoms.

Figures in this Article

Patients with acute lung injury have respiratory failure with hypoxemia (more severe in the subgroup with ARDS), bilateral pulmonary infiltrates that are not due to left atrial hypertension, and an identifiable risk factor.1 The estimated annual incidence is nearly 200,000 cases in the United States,2 and the case fatality rate is 35% to 45%.3 Current evidence suggests that survivors have persistent generalized weakness,4 reduced quality of life,5 significant cognitive impairment,6 and psychiatric morbidity7 compared with the general population.

We followed survivors of ARDS enrolled in a 5-year cohort study.4,8,9 In a previous report, we described a high prevalence (41%) of moderate to severe depressive symptoms and a lower prevalence (8% to 20%, depending on the threshold used) of memory complaints in this cohort at a median of 22 months (range, 6-48 months) after discharge from the ICU.10 Furthermore, patients reporting moderate to severe depressive symptoms were less likely to return to work than those with minimal to mild symptoms. In the current study, our objectives were to (1) evaluate the prevalence of depressive symptoms and memory complaints in the same group of survivors of ARDS at 5 years after ICU discharge; (2) identify ICU predictors of depressive symptoms and memory complaints; and (3) examine the association between depressive symptom severity and return-to-work outcomes.

Patients

The patients in this study participated in a prospective cohort study of survivors of ARDS enrolled at four University of Toronto-affiliated ICUs between May 1998 and May 2001.4,8,9 Eligible patients were aged ≥ 16 years, had a Pao2/Fio2 ratio of ≤ 200 while receiving mechanical ventilation, with a positive end-expiratory pressure of ≥ 5 cm H2O; airspace changes in all four quadrants on chest radiography; and an identifiable risk factor for ARDS. Patients were excluded if they were immobile prior to ICU admission, had a history of lung resection, or had a neurologic disease or psychiatric disorder documented in their chart. We obtained informed consent for questionnaire completion. The University Health Network Research Ethics Board approved this study (98-H015).

Survey Administration and Outcomes

We administered the Beck Depression Inventory II (BDI-II)11 and Memory Assessment Clinics Self-Rating Scale (MAC-S)1214 to patients twice: (1) by mail at a median of 22 months (range, 6-48 months) following ICU discharge10 and (2) in person at follow-up 5 years after ICU discharge. We also administered the Medical Outcomes Study 36-Item Short Form (SF-36)15 in person during annual postdischarge visits. Study personnel or family members helped to administer the instruments for patients who needed assistance (eg, translation for non-English readers). Self-reported psychiatric diagnoses after hospital discharge are reported elsewhere.9

The BDI-II instrument screens for depression using criteria consistent with the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, with higher scores (range, 0-63) indicating more severe depressive symptoms. There are two subscales measuring cognitive (nine items) and somatic-affective (12 items) symptoms,11 a factor structure validated in medical patients.16,17 Based on testing in psychiatric outpatients, symptom severity is classified as minimal (score of 0-13), mild (14-19), moderate (20-28), or severe (29-63).

The MAC-S instrument measures self-reported performance in daily memory tasks, and is divided into the ability subscale (21 items probing the ability to remember specific types of information) and frequency of occurrence subscale (24 items asking about the frequency of particular memory problems). Higher scores (range for ability, 21-105; range for frequency of occurrence, 24-120) indicate better performance.

The SF-3615,18,19 measures health-related quality of life in eight domains, including role emotional (RE) (limitations in usual role activities because of emotional problems) and general mental health (MH), with each domain scored from 0 (worst quality of life) to 100 (best). The mental component summary (MCS)20 is an aggregate score. The SF-36 has Canadian population-based norms21 and has been used in critically ill patients.22,23

Statistical Analysis

For descriptive data, we summarized nonnormally distributed continuous data using medians (first-third quartile [Q1-Q3]) and compared groups using Wilcoxon rank sum tests. Categorical data were summarized as proportions and compared using χ2 or Fisher exact tests. We compared 5-year and earlier scores using Wilcoxon signed rank test for continuous variables and exact McNemar test (because of small cell sizes) for categorized scores, including patients who contributed data at both times. We used Spearman correlation to measure the correlation between instruments (BDI-II cognitive vs somatic-affective subscales; BDI-II vs MAC-S subscales; BDI-II and subscales vs each mental health domain of SF-36). We excluded questionnaires with any missing items from all analyses but included them when describing respondent baseline characteristics.

We were interested in hypothesis-generating analyses of predictors of 5-year BDI-II and MAC-S scores, including a priori selected baseline and ICU variables. We tested each potential predictor in a univariable linear regression analysis and in a model adjusting for the earlier BDI-II or MAC-S score and time between questionnaire administrations. For the BDI-II models, we log-transformed all BDI-II scores and two predictor variables (days of mechanical ventilation and ICU stay) to ensure normally distributed residuals. We log-transformed the same two predictor variables for the MAC-S models because their distributions were skewed but did not transform the MAC-S scores. For the MAC-S adjusted analyses, we excluded cases with high residuals (n = 1 for ability subscale and n = 3 for frequency subscale) to ensure stability of the estimates of the regression coefficients.

We conducted two sensitivity analyses. The first explored the effects of missing data by including questionnaires with < 50% missing items and calculating an adjusted score based on items answered as follows: (total possible score for all items × score for items answered)/maximum possible score for items answered. The second explored the effects of including the three cases with high residuals in the MAC-S frequency regression models.

We analyzed the relationship between moderate to severe vs minimal to mild depressive symptoms (defined by 5-year BDI-II scores) and (1) return to any work and (2) return to previous position (both at 5 years). Any work included paid and unpaid work inside or outside the home. Because of small cell sizes (n < 5) in two-by-two tables, we used exact methods to generate ORs, 95% CIs, and P values. All statistical tests were two sided; we interpreted P < .05 as statistically significant. Analyses were conducted using SAS, version 9.2 (SAS Institute Inc; Cary, North Carolina) statistical software.

Study Participants

We enrolled 109 survivors of ARDS in the cohort, of whom 21 died and 24 were lost to follow-up by the time of the 5-year evaluation9 (Fig 1). Of 74 known survivors at 5 years, 64 patients were evaluated, of whom 46 (71.9%) returned the BDI-II, 47 (73.4%) returned the MAC-S, and 48 (75.0%) returned either one. A median of 41 months (Q1-Q3, 32.5-52 months) separated the two questionnaire administrations.10 Responders were similar to nonresponders (Table 1); they were relatively young (median age, 48 years; Q1-Q3, 39-61 years), and most were men (52.1%), spoke English as a first language (66.7%), and had some postsecondary education (60.5%). Responders had a high APACHE (Acute Physiology and Chronic Health Evaluation) II score,24 maximum multiple organ dysfunction score,25 maximum lung injury score,26 and persistently limited 6-min walk distance (median, 79% predicted; Q1-Q3, 60%-89% predicted).

Figure Jump LinkFigure 1. Study flow. BDI-II = Beck Depression Inventory II; MAC-S = Memory Assessment Clinics Self-Rating Scale.Grahic Jump Location
Table Graphic Jump Location
Table 1 —Characteristics of Survivors of ARDS

Data are presented as median (first-third quartile) or No. (%). The 48 responders returned either a BDI-II or MAC-S questionnaire. APACHE = Acute Physiology and Chronic Health Evaluation; BDI-II = Beck Depression Inventory II; LIS = Lung Injury Score; MAC-S = Memory Assessment Clinics Self-Rating Scale; MODS = Multiple Organ Dysfunction Score.

a 

n = 47.

b 

The LIS included the sum of the chest radiography, hypoxemia, and positive end-expiratory pressure scores, while excluding static compliance.

c 

n = 45.

d 

n = 9.

Instrument Scores and Correlations

Of the 64 survivors evaluated, 43 (67.2%), 46 (71.9%), and 38 (59.4%) patients fully completed the BDI-II, MAC-S ability subscale, and MAC-S frequency of occurrence subscale, respectively. Linear regression analyses of change in instrument scores vs months between the two administrations showed no significant associations (e-Figure 1). Histograms of scores on each instrument are presented in Figure 2.

Figure Jump LinkFigure 2. Histograms of BDI-II (n = 43), MAC-S ability subscale (n = 48), and MAC-S frequency of occurrence subscale (n = 38). Categories represent 10-point bins and only include questionnaires with no missing items. See Figure 1 legend for expansion of abbreviations.Grahic Jump Location

The median BDI-II score at 5 years was 10 (Q1-Q3, 3-18), similar to earlier scores (P = .11). Using BDI-II-defined categories, 35 respondents (81.4%) reported minimal or mild depressive symptoms and eight (18.6%) reported moderate or severe symptoms. There was no significant difference (P = .15) in the number of patients in the moderate to severe category at 5 years. Cognitive and somatic-affective subscale scores were highly correlated (Spearman correlation coefficient, 0.77; P < .001).

Scores on the MAC-S instrument were 81 (Q1-Q3, 57-92) for the ability subscale and 91.5 (Q1-Q3, 76-105) for the frequency of occurrence subscale and were unchanged from earlier scores (ability, P = .67; frequency, P = .64). Relative to a US community-based sample,12,13 4.3% of respondents in the ability subscale and none in the frequency subscale scored lower than 2 SDs below age-adjusted norms. These proportions increased to 8.7% and 15.2% (ability subscale) and 10.5% and 10.5% (frequency subscale), with cut points of 1.5 SDs and 1 SD, respectively (Table 2). BDI-II scores were moderately negatively correlated with MAC-S scores (Spearman correlation coefficient, −0.62 and −0.76 for ability and frequency of occurrence subscales, respectively; P < .0001), implying an association between increasing symptoms of depression and lower memory scores (ie, more memory complaints).

Table Graphic Jump Location
Table 2 —Depressive Symptoms and Memory Function in Survivors of ARDS

Data are presented as median (first-third quartile) or No. (%). Percentages may not sum to 100% because of rounding. The two n values in the first column refer to the earlier questionnaires (completed at a median of 22 months [range, 6-48 months] following ICU discharge for BDI-II and MAC-S10 and 2 y for SF-36) and the 5-y questionnaires, respectively. Of 64 survivors evaluated at 5 y, 48 patients answered either the BDI-II or the MAC-S; data for the earlier time point includes only patients who also completed the 5-y questionnaires. Comparisons between 5-y and earlier scores were made using Wilcoxon signed rank test (continuous variables) and exact McNemar test (categorized scores) and only included patients who contributed data at both time points. MCS = mental component summary; MH = mental health; RE = role emotional; SF-36 = Medical Outcomes Study 36-Item Short Form. See Table 1 legend for expansion of other abbreviations.

a 

At 5 y (n = 48), two patients did not answer BDI-II, and three had missing items; at the earlier administration, two did not answer, and three had missing items.

b 

Depression categories are from the BDI-II scale. The P value refers to the comparison of minimal to mild vs moderate to severe categories at 5 y vs the earlier assessment.

c 

At 5 years (n = 48), one patient did not answer, and one had a missing item; at the earlier administration, one did not answer, and five had missing items.

d 

Proportion of sample < 2, < 1.5, or < 1 SD below age-adjusted US sample mean.14 We used patients’ ages at questionnaire completion.

e 

At 5 y (n = 48), one patient did not answer, and nine had missing items; at the earlier assessment, one did not answer, and five had missing items.

f 

At 5 y (n = 48), one patient did not answer, and three, one, and four patients had missing items for RE, MH, and MCS, respectively. At the earlier assessment, eight patients did not answer, and one and two patients had missing items for RE and MCS, respectively.

g 

Proportion of sample < 2, < 1.5, or < 1 SD below age- and sex-matched Canadian sample mean.21 We used patients’ ages at questionnaire completion.

Patients in this sample had SF-36 scores of 100 (Q1-Q3, 66.7-100) for RE, 78 (Q1-Q3, 60-92) for MH, and 52 (Q1-Q3, 39-58) for MCS. Relative to a Canadian population sample,21 13.6%, 8.7%, and 14.0% scored more than 2 SDs below age-adjusted norms for these domains. These percentages rose to 27.3%, 28.3%, and 27.9%, respectively, with a cut point of 1 SD. Correlations between BDI-II (total score and cognitive and somatic-affective subscales) and SF-36 (RE, MH, and MCS) were moderately to highly negative (Table 3).

Table Graphic Jump Location
Table 3 —Agreement Between BDI-II and SF-36 Mental Health Domains

See Table 1 and 2 legends for expansion of abbreviations.

Unadjusted analyses (Tables 4, 5) demonstrated no demographic or illness severity associations with BDI-II or MAC-S scores. In adjusted analyses, higher earlier BDI-II or MAC-S scores consistently predicted higher 5-year scores. A more positive multiple organ dysfunction score slope (ie, slower recovery of organ function), longer duration of mechanical ventilation, and longer ICU stay were associated with higher BDI-II scores (ie, more severe depressive symptoms). A more positive lung injury score slope (ie, slower recovery of lung injury) had a similar effect, but statistical significance was borderline. For BDI-II, sensitivity analyses that included adjusted scores for questionnaires with missing items gave similar results. Female sex was associated with a lower MAC-S ability score (ie, more severe memory complaints) in an adjusted analysis, but the effect disappeared in a sensitivity analysis that included questionnaires with missing data. In a sensitivity analysis for MAC-S frequency of occurrence that included the three cases with high residuals, the adjusted analysis also showed a significant effect of female sex, but this effect was no longer significant if questionnaires with missing items were included.

Table Graphic Jump Location
Table 4 —Predictors of Log-Transformed BDI-II Score at Five Years in Univariable Analyses

Predictor variables refer to the index ICU admission (except for age, which is at the time of questionnaire administration). Unadjusted analyses included questionnaires with no missing items (n = 43 except for APACHE II [n = 42]). Separately for each predictor, we also included the earlier BDI-II score and the number of months between BDI-II administrations (n = 38 except for APACHE II [n = 37]) in an adjusted analysis. All BDI-II scores were log-transformed. We added 0.5 to 0 scores before taking the logarithm (n = 4 at 5 y in unadjusted analyses; n = 3 at 5 y and n = 3 for the earlier questionnaire in adjusted analyses). Positive (negative) β coefficients imply that the predictor is associated with higher (lower) log-transformed BDI-II scores. In adjusted analyses, β values for the earlier BDI-II score ranged from 0.39 to 0.46 and were highly significant (P = .003-.015); β values for the time between administrations were small and not statistically significant. In sensitivity analyses calculating adjusted scores for questionnaires with missing items, results are similar. See Table 1 and 2 legends for expansion of abbreviations.

a 

The change in MODS over time during ICU admission is expressed as the slope of the score.

b 

The change in LIS over time during ICU admission is expressed as the slope of the score.

c 

The logarithm of this variable was used because the untransformed variable had a skewed distribution.

Table Graphic Jump Location
Table 5 —Predictors of MAC-S at Five Years in Univariable Analyses

Predictor variables refer to the index ICU admission (except for age, which is at the time of questionnaire administration). Unadjusted analyses include questionnaires with complete data (n = 46 for ability subscale except for APACHE II [n = 45]; n = 38 for frequency of occurrence subscale). Separately for each predictor, we also included the earlier MAC-S score and the number of months between MAC-S administrations in an adjusted analysis. For the ability subscale, adjusted analyses (n = 39 for all except APACHE II [n = 38]) excluded one case with high residuals. For the frequency of occurrence subscale, adjusted analyses (n = 30) excluded three cases with high residuals. Positive (negative) β coefficients imply that the predictor is associated with higher (lower) MAC-S scores. In adjusted analyses, β values for the earlier MAC-S score ranged from 0.84 to 0.94 and were highly significant (all P < .0001); β values for the time between administrations were small and not statistically significant. See Table 1 and 2 legends for expansion of abbreviations.

a 

In a sensitivity analysis calculating adjusted scores for questionnaires with missing items, results were similar, but female sex was no longer significant in the adjusted analysis.

b 

The change in MODS over time during ICU admission is expressed as the slope of the score.

c 

The change in LIS over time during ICU admission is expressed as the slope of the score.

d 

The logarithm of this variable was used because the untransformed variable had a skewed distribution.

e 

In a sensitivity analysis including the three cases with high residuals, the adjusted analyses showed a significant effect of female sex (β, −11.60; SE, 4.60; P = .02). However, this effect was no longer significant if questionnaires with missing items were included by calculating adjusted scores.

Depressive Symptoms and Return to Work

Most patients returned to work at 5 years. Of 35 patients with minimal to mild depressive symptoms, 31 (88.6%) had returned to any work, and 30 (85.7%) had returned to their previous position. Of eight patients with moderate to severe symptoms of depression, five (62.5%) had returned to any work, and four (50.0%) had returned to their previous position. Patients with moderate to severe vs minimal to mild depressive symptoms tended to be less likely to return to any work (OR, 0.23; 95% CI, 0.03-1.53; P = .12) or their previous position (OR, 0.18; 95% CI, 0.03-1.01; P = .051), but the results were not statistically significant.

This study included 48 relatively young patients with high illness severity and no documented psychiatric illness who survived ARDS and answered questionnaires assessing depressive symptoms and memory complaints 5 years after ICU discharge. We found that depressive symptoms at 5 years were similar compared with ∼ 2 years after ICU discharge. The SF-36 mental health domains (MH, RE, and MCS), although correlated with BDI-II at 5 years, were completely stable between assessments and may lack sensitivity to small, but clinically important changes in mood in survivors of critical illness. Memory complaints were reported by 0% to 15.2% of patients, depending on the domain (ability or frequency of occurrence) and threshold used to define a deficit; the proportion was similar at the earlier assessment. Given the small numbers, it is difficult to draw definitive conclusions regarding depressive symptoms and return to work, but the majority returned to the workforce in some capacity, even with continuing symptoms. The most consistent predictor of scores on each instrument at 5 years was the earlier score. Adjusted analyses also found that slower resolution of multiple organ dysfunction and longer duration of ICU stay and mechanical ventilation may be associated with worse depressive symptoms at 5 years. Interestingly, slower resolution of multiple organ dysfunction and lung injury also may be associated with a shorter 6-min walk distance in these patients.9

Other studies evaluating depressive symptoms earlier after ICU discharge have reported similar findings. A systematic review reported a median point prevalence of clinically significant depression of 28% (range, 17%-43%) in four studies of survivors of ARDS, using various instruments administered within ∼ 2 years of ICU discharge.7 Factors related to subsequent severity of depressive symptoms included days in ICU, days of mechanical ventilation, and days of sedation27; surgical vs trauma or medical diagnosis28; alcohol dependence; female sex and younger age29; admission to surgical vs medical or trauma ICU and more organ dysfunction30; and higher daily ICU benzodiazepine dose.31 Hopkins et al29 found that depression in survivors of ARDS at 1 year after hospital discharge (16% of survivors), along with the presence of cognitive sequelae, predicted depression at 2 years (23% of survivors). Taken together, these observations suggest that the prevalence of depressive symptoms in survivors of ARDS remains relatively stable over time, thus justifying a program of active and ongoing psychiatric surveillance.

Studies3239 performing formal neurocognitive testing (assessing mental processing speed, memory, attention, problem-solving [executive function], intellectual function, and visual-spatial ability) after ARDS generally have found a higher prevalence of impairments compared with the prevalence of memory complaints that we reported earlier10 and in the present study. As we noted previously,10 this discordance has been noted in other populations and may arise from the presence of more severe objective memory impairment than patients perceived or poor correlation of self-reported memory complaints with objective testing.4049 Others have found no consistent associations among baseline clinical variables, illness severity, and postdischarge neurocognitive dysfunction.6

To our knowledge, no other study has measured self-reported depressive symptoms or memory complaints using validated instruments in critically ill patients 5 years after ICU discharge in a well-characterized inception cohort. The major strength of the present study is the reliable ascertainment of exposure variables (baseline status and selected ICU treatment and physiologic variables) and outcomes (symptoms of depression and memory complaints). Our study also has weaknesses. Of the 74 survivors of ARDS at 5 years, 10 (13.5%) were lost to follow-up and 64 (86.5%) were evaluated in the clinic, but only 48 (64.9%) attempted to complete BDI-II or MAC-S questionnaires. Because the sample size was small, the power to detect clinically important differences between responders and nonresponders was limited. In other studies evaluating quality of life ≥ 5 years after ICU discharge, rates of follow-up of known survivors were 48.0%,50 62.5%,51 66.0%,52 and 91.1%.53 The present nonresponse rate of approximately one-third of survivors is, therefore, similar to several other studies. However, subsequent biases in instrument responses are unpredictable because the direction of confounding due to clinically dissimilar baseline characteristics between responders and nonresponders may be unknown or inconsistent. In addition, we likely had low power to detect clinically important associations between questionnaire scores and the variables examined.

Other weaknesses of the study are unchanged.10 The sample consists of relatively young and employed patients with few comorbidities at the time of critical illness, and the findings may not be generalizable to older and sicker survivors of critical illness who may be expected to have more severe long-term memory impairments and mood disorders. We did not collect data on other possible predictors of BDI-II and MAC-S, such as medications, hypoxemia,32,35,36 hypoglycemia,54 or environmental issues in the ICU, or intervening events between ICU discharge and questionnaire administration. We excluded patients with psychiatric disorders documented in their medical chart but did not formally assess premorbid cognitive function or mood. We did not screen for other mood disorders, delirium, or dementia or conduct standardized psychiatric interviews. Finally, the relationship between the subjective instruments we used and objective neurocognitive testing and diagnosis of major depressive disorder is unclear and requires further research.

In summary, patients who survive ARDS to 5 years after discharge have similar scores on instruments assessing depressive symptoms and self-reported memory complaints compared with ∼ 2 years after ICU discharge. Future studies should evaluate these long-term trends in different populations along with concurrent postdischarge mental health utilization (medication use, visits to mental health providers, and hospitalization), which would enhance understanding of the trajectory of these symptoms. Screening for depressive symptoms should use specific validated instruments rather than the mental health domains of a general health-related quality-of-life instrument, such as the SF-36. Descriptions of these phenomena and their impact on quality of life in observational studies has grown considerably in contrast to the relative paucity of data from randomized trials of interventions to improve postdischarge quality of life.55 Therefore, development and evaluation of feasible, reliable, and valid methods to screen survivors of ARDS for psychiatric conditions and provide prompt referral to mental health expertise is required.

Author contributions: Dr Adhikari had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Dr Adhikari: contributed the design of the analyses, interpretation of the data, and drafting and revision of the manuscript.

Dr Tansey: contributed to the data collection and revision of the manuscript.

Dr McAndrews: contributed to the study design and revision of the manuscript.

Ms Matté: contributed to the data collection, data entry, and revision of the manuscript.

Dr Pinto: contributed to the data analysis and revision of the manuscript.

Dr Cheung: contributed to the study design and revision of the manuscript.

Ms Diaz-Granados: contributed to the study design and revision of the manuscript.

Dr Herridge: contributed to the study design and conception, obtaining of funding, supervision, data collection and interpretation, and revision of the manuscript.

Financial/nonfinancial disclosures: The authors have reported to CHEST that no potential conflicts of interest exist with any companies/organizations whose products or services may be discussed in this article.

Role of sponsors: The sponsors had no role in the design of the study, the collection and analysis of the data, or in the preparation of the manuscript.

Additional information: The e-Figure can be found in the Online Supplement at http://chestjournal.chestpubs.org/content/140/6/1484/suppl/DC1.

Author contributions: Dr Adhikari had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Dr Adhikari: contributed the design of the analyses, interpretation of the data, and drafting and revision of the manuscript.

Dr Tansey: contributed to the data collection and revision of the manuscript.

Dr McAndrews: contributed to the study design and revision of the manuscript.

Ms Matté: contributed to the data collection, data entry, and revision of the manuscript.

Dr Pinto: contributed to the data analysis and revision of the manuscript.

Dr Cheung: contributed to the study design and revision of the manuscript.

Ms Diaz-Granados: contributed to the study design and revision of the manuscript.

Dr Herridge: contributed to the study design and conception, obtaining of funding, supervision, data collection and interpretation, and revision of the manuscript.

Financial/nonfinancial disclosures: The authors have reported to CHEST that no potential conflicts of interest exist with any companies/organizations whose products or services may be discussed in this article.

Role of sponsors: The sponsors had no role in the design of the study, the collection and analysis of the data, or in the preparation of the manuscript.

Additional information: The e-Figure can be found in the Online Supplement at http://chestjournal.chestpubs.org/content/140/6/1484/suppl/DC1.

BDI-II

Beck Depression Inventory-II

MAC-S

Memory Assessment Clinics Self-Rating Scale

MCS

mental component summary

MH

mental health

Q1-Q3

first-third quartile

RE

role emotional

SF-36

Medical Outcomes Study 36-Item Short Form

Bernard GR, Artigas A, Brigham KL, et al. The American-European Consensus Conference on ARDS. Definitions, mechanisms, relevant outcomes, and clinical trial coordination. Am J Respir Crit Care Med. 1994;1493 pt 1:818-824 [PubMed]
 
Rubenfeld GD, Caldwell E, Peabody E, et al. Incidence and outcomes of acute lung injury. N Engl J Med. 2005;35316:1685-1693 [CrossRef]
 
Phua J, Badia JR, Adhikari NK, et al. Has mortality from acute respiratory distress syndrome decreased over time? a systematic review. Am J Respir Crit Care Med. 2009;1793:220-227 [CrossRef]
 
Herridge MS, Cheung AM, Tansey CM, et al; Canadian Critical Care Trials Group Canadian Critical Care Trials Group One-year outcomes in survivors of the acute respiratory distress syndrome. N Engl J Med. 2003;3488:683-693 [CrossRef]
 
Dowdy DW, Eid MP, Dennison CR, et al. Quality of life after acute respiratory distress syndrome: a meta-analysis. Intensive Care Med. 2006;328:1115-1124 [CrossRef]
 
Hopkins RO, Jackson JC. Long-term neurocognitive function after critical illness. Chest. 2006;1303:869-878 [CrossRef]
 
Davydow DS, Desai SV, Needham DM, Bienvenu OJ. Psychiatric morbidity in survivors of the acute respiratory distress syndrome: a systematic review. Psychosom Med. 2008;704:512-519 [CrossRef]
 
Cheung AM, Tansey CM, Tomlinson G, et al. Two-year outcomes, health care use, and costs of survivors of acute respiratory distress syndrome. Am J Respir Crit Care Med. 2006;1745:538-544 [CrossRef]
 
Herridge MS, Tansey CM, Matté A, et al; Canadian Critical Care Trials Group Canadian Critical Care Trials Group Functional disability 5 years after acute respiratory distress syndrome. N Engl J Med. 2011;36414:1293-1304 [CrossRef]
 
Adhikari NK, McAndrews MP, Tansey CM, et al. Self-reported symptoms of depression and memory dysfunction in survivors of ARDS. Chest. 2009;1353:678-687 [CrossRef]
 
Beck AT, Steer RA, Brown GK. Manual for the Beck Depression Inventory-II. 1996; San Antonio, TX Psychological Corporation:1-38
 
Winterling D, Crook T, Salama M, et al;Bès A, Cahn J, Hoyer S, et al. A self-rating scale for assessing memory loss. Senile Dementias: Early Detection. 1986; London, England and Paris, France John Libbey Eurotext:482-487
 
Crook TH III, Larrabee GJ. A self-rating scale for evaluating memory in everyday life. Psychol Aging. 1990;51:48-57 [CrossRef]
 
Crook TH, Larrabee GJ. Normative data on a self-rating scale for evaluating memory in everyday life. Arch Clin Neuropsychol. 1992;71:41-51
 
Ware JE Jr, Sherbourne CD. The MOS 36-item short-form health survey (SF-36). I. Conceptual framework and item selection. Med Care. 1992;306:473-483 [CrossRef]
 
Arnau RC, Meagher MW, Norris MP, Bramson R. Psychometric evaluation of the Beck Depression Inventory-II with primary care medical patients. Health Psychol. 2001;202:112-119 [CrossRef]
 
Viljoen JL, Iverson GL, Griffiths S, Woodward TS. Factor structure of the Beck Depression Inventory-II in a medical outpatient sample. J Clin Psychol Med Settings. 2003;104:289-291 [CrossRef]
 
McHorney CA, Ware JE Jr, Raczek AE. The MOS 36-Item Short-Form Health Survey (SF-36): II. Psychometric and clinical tests of validity in measuring physical and mental health constructs. Med Care. 1993;313:247-263 [CrossRef]
 
McHorney CA, Ware JE Jr, Lu JF, Sherbourne CD. The MOS 36-item Short-Form Health Survey (SF-36): III. Tests of data quality, scaling assumptions, and reliability across diverse patient groups. Med Care. 1994;321:40-66 [CrossRef]
 
Ware JE Jr, Kosinski M, Bayliss MS, McHorney CA, Rogers WH, Raczek A. Comparison of methods for the scoring and statistical analysis of SF-36 health profile and summary measures: summary of results from the Medical Outcomes Study. Med Care. 1995;33suppl 4:AS264-AS279 [CrossRef]
 
Hopman WM, Towheed T, Anastassiades T, et al; Canadian Multicentre Osteoporosis Study Research Group Canadian Multicentre Osteoporosis Study Research Group Canadian normative data for the SF-36 health survey. CMAJ. 2000;1633:265-271
 
Dowdy DW, Eid MP, Sedrakyan A, et al. Quality of life in adult survivors of critical illness: a systematic review of the literature. Intensive Care Med. 2005;315:611-620 [CrossRef]
 
Winters BD, Eberlein M, Leung J, Needham DM, Pronovost PJ, Sevransky JE. Long-term mortality and quality of life in sepsis: a systematic review. Crit Care Med. 2010;385:1276-1283
 
Knaus WA, Draper EA, Wagner DP, Zimmerman JE. APACHE II: a severity of disease classification system. Crit Care Med. 1985;1310:818-829 [CrossRef]
 
Marshall JC, Cook DJ, Christou NV, Bernard GR, Sprung CL, Sibbald WJ. Multiple organ dysfunction score: a reliable descriptor of a complex clinical outcome. Crit Care Med. 1995;2310:1638-1652 [CrossRef]
 
Murray JF, Matthay MA, Luce JM, Flick MR. An expanded definition of the adult respiratory distress syndrome [published correction appears inAm Rev Respir Dis. 1989;139(4):1065]. Am Rev Respir Dis. 1988;1383:720-723
 
Nelson BJ, Weinert CR, Bury CL, Marinelli WA, Gross CR. Intensive care unit drug use and subsequent quality of life in acute lung injury patients. Crit Care Med. 2000;2811:3626-3630 [CrossRef]
 
Myhren H, Ekeberg O, Tøien K, Karlsson S, Stokland O. Posttraumatic stress, anxiety and depression symptoms in patients during the first year post intensive care unit discharge. Crit Care. 2010;141:R14 [CrossRef]
 
Hopkins RO, Key CW, Suchyta MR, Weaver LK, Orme JF Jr. Risk factors for depression and anxiety in survivors of acute respiratory distress syndrome. Gen Hosp Psychiatry. 2010;322:147-155 [CrossRef]
 
Vincent JL, Moreno R, Takala J, et al. The SOFA (Sepsis-related Organ Failure Assessment) score to describe organ dysfunction/failure. On behalf of the Working Group on Sepsis-Related Problems of the European Society of Intensive Care Medicine. Intensive Care Med. 1996;227:707-710 [CrossRef]
 
Dowdy DW, Bienvenu OJ, Dinglas VD, et al. Are intensive care factors associated with depressive symptoms 6 months after acute lung injury? Crit Care Med. 2009;375:1702-1707 [CrossRef]
 
Hopkins RO, Weaver LK, Pope D, Orme JF, Bigler ED, Larson-LOHR V. Neuropsychological sequelae and impaired health status in survivors of severe acute respiratory distress syndrome. Am J Respir Crit Care Med. 1999;1601:50-56
 
Rothenhäusler HB, Ehrentraut S, Stoll C, Schelling G, Kapfhammer HP. The relationship between cognitive performance and employment and health status in long-term survivors of the acute respiratory distress syndrome: results of an exploratory study. Gen Hosp Psychiatry. 2001;232:90-96 [CrossRef]
 
Jackson JC, Hart RP, Gordon SM, et al. Six-month neuropsychological outcome of medical intensive care unit patients. Crit Care Med. 2003;314:1226-1234 [CrossRef]
 
Hopkins RO, Weaver LK, Chan KJ, Orme JF Jr. Quality of life, emotional, and cognitive function following acute respiratory distress syndrome. J Int Neuropsychol Soc. 2004;107:1005-1017
 
Hopkins RO, Weaver LK, Collingridge D, Parkinson RB, Chan KJ, Orme JF Jr. Two-year cognitive, emotional, and quality-of-life outcomes in acute respiratory distress syndrome. Am J Respir Crit Care Med. 2005;1714:340-347 [CrossRef]
 
Hopkins RO, Jackson JC, Wallace CJ. Neurocognitive impairments in ICU patients with prolonged mechanical ventilation [abstract]. J Int Neuropsychol Soc. 2005;11supp 1:60
 
Jones C, Griffiths RD, Slater T, Benjamin KS, Wilson S. Significant cognitive dysfunction in non-delirious patients identified during and persisting following critical illness. Intensive Care Med. 2006;326:923-926 [CrossRef]
 
Mikkelsen ME, Shull WH, Biester RC, et al. Cognitive, mood and quality of life impairments in a select population of ARDS survivors. Respirology. 2009;141:76-82 [CrossRef]
 
Baños JH, LaGory J, Sawrie S, et al. Self-report of cognitive abilities in temporal lobe epilepsy: cognitive, psychosocial, and emotional factors. Epilepsy Behav. 2004;54:575-579 [CrossRef]
 
Carter SL, Rourke SB, Murji S, Shore D, Rourke BP. Cognitive complaints, depression, medical symptoms, and their association with neuropsychological functioning in HIV infection: a structural equation model analysis. Neuropsychology. 2003;173:410-419 [CrossRef]
 
Duits A, Munnecom T, van Heugten C, van Oostenbrugge RJ. Cognitive complaints in the early phase after stroke are not indicative of cognitive impairment. J Neurol Neurosurg Psychiatry. 2008;792:143-146 [CrossRef]
 
Hilsabeck RC, Hassanein TI, Carlson MD, Ziegler EA, Perry W. Cognitive functioning and psychiatric symptomatology in patients with chronic hepatitis C. J Int Neuropsychol Soc. 2003;96:847-854 [CrossRef]
 
Booth-Jones M, Jacobsen PB, Ransom S, Soety E. Characteristics and correlates of cognitive functioning following bone marrow transplantation. Bone Marrow Transplant. 2005;368:695-702 [CrossRef]
 
Klepstad P, Hilton P, Moen J, Fougner B, Borchgrevink PC, Kaasa S. Self-reports are not related to objective assessments of cognitive function and sedation in patients with cancer pain admitted to a palliative care unit. Palliat Med. 2002;166:513-519 [CrossRef]
 
Mitchell AJ. The clinical significance of subjective memory complaints in the diagnosis of mild cognitive impairment and dementia: a meta-analysis. Int J Geriatr Psychiatry. 2008;2311:1191-1202 [CrossRef]
 
Vogel A, Elberling TV, Hørding M, et al. Affective symptoms and cognitive functions in the acute phase of Graves’ thyrotoxicosis. Psychoneuroendocrinology. 2007;321:36-43 [CrossRef]
 
Reid LM, Maclullich AM. Subjective memory complaints and cognitive impairment in older people. Dement Geriatr Cogn Disord. 2006;225-6:471-485 [CrossRef]
 
Robinson JP, Burwinkle T, Turk DC. Perceived and actual memory, concentration, and attention problems after whiplash-associated disorders (grades I and II): prevalence and predictors. Arch Phys Med Rehabil. 2007;886:774-779 [CrossRef]
 
Cuthbertson BH, Roughton S, Jenkinson D, Maclennan G, Vale L. Quality of life in the five years after intensive care: a cohort study. Crit Care. 2010;141:R6 [CrossRef]
 
Schelling G, Stoll C, Vogelmeier C, et al. Pulmonary function and health-related quality of life in a sample of long-term survivors of the acute respiratory distress syndrome. Intensive Care Med. 2000;269:1304-1311 [CrossRef]
 
Kaarlola A, Pettilä V, Kekki P. Quality of life six years after intensive care. Intensive Care Med. 2003;298:1294-1299 [CrossRef]
 
Graf J, Wagner J, Graf C, Koch KC, Janssens U. Five-year survival, quality of life, and individual costs of 303 consecutive medical intensive care patients—a cost-utility analysis. Crit Care Med. 2005;333:547-555 [CrossRef]
 
Dowdy DW, Dinglas V, Mendez-Tellez PA, et al. Intensive care unit hypoglycemia predicts depression during early recovery from acute lung injury. Crit Care Med. 2008;3610:2726-2733 [CrossRef]
 
Cuthbertson BH, Rattray J, Campbell MK, et al; PRaCTICaL study group PRaCTICaL study group The PRaCTICaL study of nurse led, intensive care follow-up programmes for improving long term outcomes from critical illness: a pragmatic randomised controlled trial. BMJ. 2009;339:b3723 [CrossRef]
 

Figures

Figure Jump LinkFigure 1. Study flow. BDI-II = Beck Depression Inventory II; MAC-S = Memory Assessment Clinics Self-Rating Scale.Grahic Jump Location
Figure Jump LinkFigure 2. Histograms of BDI-II (n = 43), MAC-S ability subscale (n = 48), and MAC-S frequency of occurrence subscale (n = 38). Categories represent 10-point bins and only include questionnaires with no missing items. See Figure 1 legend for expansion of abbreviations.Grahic Jump Location

Tables

Table Graphic Jump Location
Table 1 —Characteristics of Survivors of ARDS

Data are presented as median (first-third quartile) or No. (%). The 48 responders returned either a BDI-II or MAC-S questionnaire. APACHE = Acute Physiology and Chronic Health Evaluation; BDI-II = Beck Depression Inventory II; LIS = Lung Injury Score; MAC-S = Memory Assessment Clinics Self-Rating Scale; MODS = Multiple Organ Dysfunction Score.

a 

n = 47.

b 

The LIS included the sum of the chest radiography, hypoxemia, and positive end-expiratory pressure scores, while excluding static compliance.

c 

n = 45.

d 

n = 9.

Table Graphic Jump Location
Table 2 —Depressive Symptoms and Memory Function in Survivors of ARDS

Data are presented as median (first-third quartile) or No. (%). Percentages may not sum to 100% because of rounding. The two n values in the first column refer to the earlier questionnaires (completed at a median of 22 months [range, 6-48 months] following ICU discharge for BDI-II and MAC-S10 and 2 y for SF-36) and the 5-y questionnaires, respectively. Of 64 survivors evaluated at 5 y, 48 patients answered either the BDI-II or the MAC-S; data for the earlier time point includes only patients who also completed the 5-y questionnaires. Comparisons between 5-y and earlier scores were made using Wilcoxon signed rank test (continuous variables) and exact McNemar test (categorized scores) and only included patients who contributed data at both time points. MCS = mental component summary; MH = mental health; RE = role emotional; SF-36 = Medical Outcomes Study 36-Item Short Form. See Table 1 legend for expansion of other abbreviations.

a 

At 5 y (n = 48), two patients did not answer BDI-II, and three had missing items; at the earlier administration, two did not answer, and three had missing items.

b 

Depression categories are from the BDI-II scale. The P value refers to the comparison of minimal to mild vs moderate to severe categories at 5 y vs the earlier assessment.

c 

At 5 years (n = 48), one patient did not answer, and one had a missing item; at the earlier administration, one did not answer, and five had missing items.

d 

Proportion of sample < 2, < 1.5, or < 1 SD below age-adjusted US sample mean.14 We used patients’ ages at questionnaire completion.

e 

At 5 y (n = 48), one patient did not answer, and nine had missing items; at the earlier assessment, one did not answer, and five had missing items.

f 

At 5 y (n = 48), one patient did not answer, and three, one, and four patients had missing items for RE, MH, and MCS, respectively. At the earlier assessment, eight patients did not answer, and one and two patients had missing items for RE and MCS, respectively.

g 

Proportion of sample < 2, < 1.5, or < 1 SD below age- and sex-matched Canadian sample mean.21 We used patients’ ages at questionnaire completion.

Table Graphic Jump Location
Table 3 —Agreement Between BDI-II and SF-36 Mental Health Domains

See Table 1 and 2 legends for expansion of abbreviations.

Table Graphic Jump Location
Table 4 —Predictors of Log-Transformed BDI-II Score at Five Years in Univariable Analyses

Predictor variables refer to the index ICU admission (except for age, which is at the time of questionnaire administration). Unadjusted analyses included questionnaires with no missing items (n = 43 except for APACHE II [n = 42]). Separately for each predictor, we also included the earlier BDI-II score and the number of months between BDI-II administrations (n = 38 except for APACHE II [n = 37]) in an adjusted analysis. All BDI-II scores were log-transformed. We added 0.5 to 0 scores before taking the logarithm (n = 4 at 5 y in unadjusted analyses; n = 3 at 5 y and n = 3 for the earlier questionnaire in adjusted analyses). Positive (negative) β coefficients imply that the predictor is associated with higher (lower) log-transformed BDI-II scores. In adjusted analyses, β values for the earlier BDI-II score ranged from 0.39 to 0.46 and were highly significant (P = .003-.015); β values for the time between administrations were small and not statistically significant. In sensitivity analyses calculating adjusted scores for questionnaires with missing items, results are similar. See Table 1 and 2 legends for expansion of abbreviations.

a 

The change in MODS over time during ICU admission is expressed as the slope of the score.

b 

The change in LIS over time during ICU admission is expressed as the slope of the score.

c 

The logarithm of this variable was used because the untransformed variable had a skewed distribution.

Table Graphic Jump Location
Table 5 —Predictors of MAC-S at Five Years in Univariable Analyses

Predictor variables refer to the index ICU admission (except for age, which is at the time of questionnaire administration). Unadjusted analyses include questionnaires with complete data (n = 46 for ability subscale except for APACHE II [n = 45]; n = 38 for frequency of occurrence subscale). Separately for each predictor, we also included the earlier MAC-S score and the number of months between MAC-S administrations in an adjusted analysis. For the ability subscale, adjusted analyses (n = 39 for all except APACHE II [n = 38]) excluded one case with high residuals. For the frequency of occurrence subscale, adjusted analyses (n = 30) excluded three cases with high residuals. Positive (negative) β coefficients imply that the predictor is associated with higher (lower) MAC-S scores. In adjusted analyses, β values for the earlier MAC-S score ranged from 0.84 to 0.94 and were highly significant (all P < .0001); β values for the time between administrations were small and not statistically significant. See Table 1 and 2 legends for expansion of abbreviations.

a 

In a sensitivity analysis calculating adjusted scores for questionnaires with missing items, results were similar, but female sex was no longer significant in the adjusted analysis.

b 

The change in MODS over time during ICU admission is expressed as the slope of the score.

c 

The change in LIS over time during ICU admission is expressed as the slope of the score.

d 

The logarithm of this variable was used because the untransformed variable had a skewed distribution.

e 

In a sensitivity analysis including the three cases with high residuals, the adjusted analyses showed a significant effect of female sex (β, −11.60; SE, 4.60; P = .02). However, this effect was no longer significant if questionnaires with missing items were included by calculating adjusted scores.

References

Bernard GR, Artigas A, Brigham KL, et al. The American-European Consensus Conference on ARDS. Definitions, mechanisms, relevant outcomes, and clinical trial coordination. Am J Respir Crit Care Med. 1994;1493 pt 1:818-824 [PubMed]
 
Rubenfeld GD, Caldwell E, Peabody E, et al. Incidence and outcomes of acute lung injury. N Engl J Med. 2005;35316:1685-1693 [CrossRef]
 
Phua J, Badia JR, Adhikari NK, et al. Has mortality from acute respiratory distress syndrome decreased over time? a systematic review. Am J Respir Crit Care Med. 2009;1793:220-227 [CrossRef]
 
Herridge MS, Cheung AM, Tansey CM, et al; Canadian Critical Care Trials Group Canadian Critical Care Trials Group One-year outcomes in survivors of the acute respiratory distress syndrome. N Engl J Med. 2003;3488:683-693 [CrossRef]
 
Dowdy DW, Eid MP, Dennison CR, et al. Quality of life after acute respiratory distress syndrome: a meta-analysis. Intensive Care Med. 2006;328:1115-1124 [CrossRef]
 
Hopkins RO, Jackson JC. Long-term neurocognitive function after critical illness. Chest. 2006;1303:869-878 [CrossRef]
 
Davydow DS, Desai SV, Needham DM, Bienvenu OJ. Psychiatric morbidity in survivors of the acute respiratory distress syndrome: a systematic review. Psychosom Med. 2008;704:512-519 [CrossRef]
 
Cheung AM, Tansey CM, Tomlinson G, et al. Two-year outcomes, health care use, and costs of survivors of acute respiratory distress syndrome. Am J Respir Crit Care Med. 2006;1745:538-544 [CrossRef]
 
Herridge MS, Tansey CM, Matté A, et al; Canadian Critical Care Trials Group Canadian Critical Care Trials Group Functional disability 5 years after acute respiratory distress syndrome. N Engl J Med. 2011;36414:1293-1304 [CrossRef]
 
Adhikari NK, McAndrews MP, Tansey CM, et al. Self-reported symptoms of depression and memory dysfunction in survivors of ARDS. Chest. 2009;1353:678-687 [CrossRef]
 
Beck AT, Steer RA, Brown GK. Manual for the Beck Depression Inventory-II. 1996; San Antonio, TX Psychological Corporation:1-38
 
Winterling D, Crook T, Salama M, et al;Bès A, Cahn J, Hoyer S, et al. A self-rating scale for assessing memory loss. Senile Dementias: Early Detection. 1986; London, England and Paris, France John Libbey Eurotext:482-487
 
Crook TH III, Larrabee GJ. A self-rating scale for evaluating memory in everyday life. Psychol Aging. 1990;51:48-57 [CrossRef]
 
Crook TH, Larrabee GJ. Normative data on a self-rating scale for evaluating memory in everyday life. Arch Clin Neuropsychol. 1992;71:41-51
 
Ware JE Jr, Sherbourne CD. The MOS 36-item short-form health survey (SF-36). I. Conceptual framework and item selection. Med Care. 1992;306:473-483 [CrossRef]
 
Arnau RC, Meagher MW, Norris MP, Bramson R. Psychometric evaluation of the Beck Depression Inventory-II with primary care medical patients. Health Psychol. 2001;202:112-119 [CrossRef]
 
Viljoen JL, Iverson GL, Griffiths S, Woodward TS. Factor structure of the Beck Depression Inventory-II in a medical outpatient sample. J Clin Psychol Med Settings. 2003;104:289-291 [CrossRef]
 
McHorney CA, Ware JE Jr, Raczek AE. The MOS 36-Item Short-Form Health Survey (SF-36): II. Psychometric and clinical tests of validity in measuring physical and mental health constructs. Med Care. 1993;313:247-263 [CrossRef]
 
McHorney CA, Ware JE Jr, Lu JF, Sherbourne CD. The MOS 36-item Short-Form Health Survey (SF-36): III. Tests of data quality, scaling assumptions, and reliability across diverse patient groups. Med Care. 1994;321:40-66 [CrossRef]
 
Ware JE Jr, Kosinski M, Bayliss MS, McHorney CA, Rogers WH, Raczek A. Comparison of methods for the scoring and statistical analysis of SF-36 health profile and summary measures: summary of results from the Medical Outcomes Study. Med Care. 1995;33suppl 4:AS264-AS279 [CrossRef]
 
Hopman WM, Towheed T, Anastassiades T, et al; Canadian Multicentre Osteoporosis Study Research Group Canadian Multicentre Osteoporosis Study Research Group Canadian normative data for the SF-36 health survey. CMAJ. 2000;1633:265-271
 
Dowdy DW, Eid MP, Sedrakyan A, et al. Quality of life in adult survivors of critical illness: a systematic review of the literature. Intensive Care Med. 2005;315:611-620 [CrossRef]
 
Winters BD, Eberlein M, Leung J, Needham DM, Pronovost PJ, Sevransky JE. Long-term mortality and quality of life in sepsis: a systematic review. Crit Care Med. 2010;385:1276-1283
 
Knaus WA, Draper EA, Wagner DP, Zimmerman JE. APACHE II: a severity of disease classification system. Crit Care Med. 1985;1310:818-829 [CrossRef]
 
Marshall JC, Cook DJ, Christou NV, Bernard GR, Sprung CL, Sibbald WJ. Multiple organ dysfunction score: a reliable descriptor of a complex clinical outcome. Crit Care Med. 1995;2310:1638-1652 [CrossRef]
 
Murray JF, Matthay MA, Luce JM, Flick MR. An expanded definition of the adult respiratory distress syndrome [published correction appears inAm Rev Respir Dis. 1989;139(4):1065]. Am Rev Respir Dis. 1988;1383:720-723
 
Nelson BJ, Weinert CR, Bury CL, Marinelli WA, Gross CR. Intensive care unit drug use and subsequent quality of life in acute lung injury patients. Crit Care Med. 2000;2811:3626-3630 [CrossRef]
 
Myhren H, Ekeberg O, Tøien K, Karlsson S, Stokland O. Posttraumatic stress, anxiety and depression symptoms in patients during the first year post intensive care unit discharge. Crit Care. 2010;141:R14 [CrossRef]
 
Hopkins RO, Key CW, Suchyta MR, Weaver LK, Orme JF Jr. Risk factors for depression and anxiety in survivors of acute respiratory distress syndrome. Gen Hosp Psychiatry. 2010;322:147-155 [CrossRef]
 
Vincent JL, Moreno R, Takala J, et al. The SOFA (Sepsis-related Organ Failure Assessment) score to describe organ dysfunction/failure. On behalf of the Working Group on Sepsis-Related Problems of the European Society of Intensive Care Medicine. Intensive Care Med. 1996;227:707-710 [CrossRef]
 
Dowdy DW, Bienvenu OJ, Dinglas VD, et al. Are intensive care factors associated with depressive symptoms 6 months after acute lung injury? Crit Care Med. 2009;375:1702-1707 [CrossRef]
 
Hopkins RO, Weaver LK, Pope D, Orme JF, Bigler ED, Larson-LOHR V. Neuropsychological sequelae and impaired health status in survivors of severe acute respiratory distress syndrome. Am J Respir Crit Care Med. 1999;1601:50-56
 
Rothenhäusler HB, Ehrentraut S, Stoll C, Schelling G, Kapfhammer HP. The relationship between cognitive performance and employment and health status in long-term survivors of the acute respiratory distress syndrome: results of an exploratory study. Gen Hosp Psychiatry. 2001;232:90-96 [CrossRef]
 
Jackson JC, Hart RP, Gordon SM, et al. Six-month neuropsychological outcome of medical intensive care unit patients. Crit Care Med. 2003;314:1226-1234 [CrossRef]
 
Hopkins RO, Weaver LK, Chan KJ, Orme JF Jr. Quality of life, emotional, and cognitive function following acute respiratory distress syndrome. J Int Neuropsychol Soc. 2004;107:1005-1017
 
Hopkins RO, Weaver LK, Collingridge D, Parkinson RB, Chan KJ, Orme JF Jr. Two-year cognitive, emotional, and quality-of-life outcomes in acute respiratory distress syndrome. Am J Respir Crit Care Med. 2005;1714:340-347 [CrossRef]
 
Hopkins RO, Jackson JC, Wallace CJ. Neurocognitive impairments in ICU patients with prolonged mechanical ventilation [abstract]. J Int Neuropsychol Soc. 2005;11supp 1:60
 
Jones C, Griffiths RD, Slater T, Benjamin KS, Wilson S. Significant cognitive dysfunction in non-delirious patients identified during and persisting following critical illness. Intensive Care Med. 2006;326:923-926 [CrossRef]
 
Mikkelsen ME, Shull WH, Biester RC, et al. Cognitive, mood and quality of life impairments in a select population of ARDS survivors. Respirology. 2009;141:76-82 [CrossRef]
 
Baños JH, LaGory J, Sawrie S, et al. Self-report of cognitive abilities in temporal lobe epilepsy: cognitive, psychosocial, and emotional factors. Epilepsy Behav. 2004;54:575-579 [CrossRef]
 
Carter SL, Rourke SB, Murji S, Shore D, Rourke BP. Cognitive complaints, depression, medical symptoms, and their association with neuropsychological functioning in HIV infection: a structural equation model analysis. Neuropsychology. 2003;173:410-419 [CrossRef]
 
Duits A, Munnecom T, van Heugten C, van Oostenbrugge RJ. Cognitive complaints in the early phase after stroke are not indicative of cognitive impairment. J Neurol Neurosurg Psychiatry. 2008;792:143-146 [CrossRef]
 
Hilsabeck RC, Hassanein TI, Carlson MD, Ziegler EA, Perry W. Cognitive functioning and psychiatric symptomatology in patients with chronic hepatitis C. J Int Neuropsychol Soc. 2003;96:847-854 [CrossRef]
 
Booth-Jones M, Jacobsen PB, Ransom S, Soety E. Characteristics and correlates of cognitive functioning following bone marrow transplantation. Bone Marrow Transplant. 2005;368:695-702 [CrossRef]
 
Klepstad P, Hilton P, Moen J, Fougner B, Borchgrevink PC, Kaasa S. Self-reports are not related to objective assessments of cognitive function and sedation in patients with cancer pain admitted to a palliative care unit. Palliat Med. 2002;166:513-519 [CrossRef]
 
Mitchell AJ. The clinical significance of subjective memory complaints in the diagnosis of mild cognitive impairment and dementia: a meta-analysis. Int J Geriatr Psychiatry. 2008;2311:1191-1202 [CrossRef]
 
Vogel A, Elberling TV, Hørding M, et al. Affective symptoms and cognitive functions in the acute phase of Graves’ thyrotoxicosis. Psychoneuroendocrinology. 2007;321:36-43 [CrossRef]
 
Reid LM, Maclullich AM. Subjective memory complaints and cognitive impairment in older people. Dement Geriatr Cogn Disord. 2006;225-6:471-485 [CrossRef]
 
Robinson JP, Burwinkle T, Turk DC. Perceived and actual memory, concentration, and attention problems after whiplash-associated disorders (grades I and II): prevalence and predictors. Arch Phys Med Rehabil. 2007;886:774-779 [CrossRef]
 
Cuthbertson BH, Roughton S, Jenkinson D, Maclennan G, Vale L. Quality of life in the five years after intensive care: a cohort study. Crit Care. 2010;141:R6 [CrossRef]
 
Schelling G, Stoll C, Vogelmeier C, et al. Pulmonary function and health-related quality of life in a sample of long-term survivors of the acute respiratory distress syndrome. Intensive Care Med. 2000;269:1304-1311 [CrossRef]
 
Kaarlola A, Pettilä V, Kekki P. Quality of life six years after intensive care. Intensive Care Med. 2003;298:1294-1299 [CrossRef]
 
Graf J, Wagner J, Graf C, Koch KC, Janssens U. Five-year survival, quality of life, and individual costs of 303 consecutive medical intensive care patients—a cost-utility analysis. Crit Care Med. 2005;333:547-555 [CrossRef]
 
Dowdy DW, Dinglas V, Mendez-Tellez PA, et al. Intensive care unit hypoglycemia predicts depression during early recovery from acute lung injury. Crit Care Med. 2008;3610:2726-2733 [CrossRef]
 
Cuthbertson BH, Rattray J, Campbell MK, et al; PRaCTICaL study group PRaCTICaL study group The PRaCTICaL study of nurse led, intensive care follow-up programmes for improving long term outcomes from critical illness: a pragmatic randomised controlled trial. BMJ. 2009;339:b3723 [CrossRef]
 
NOTE:
Citing articles are presented as examples only. In non-demo SCM6 implementation, integration with CrossRef’s "Cited By" API will populate this tab (http://www.crossref.org/citedby.html).
Supporting Data
Data Supplement

Some tools below are only available to our subscribers or users with an online account.

Related Content

Customize your page view by dragging & repositioning the boxes below.

CHEST Journal Articles
PubMed Articles
Guidelines
Detection and assessment of late life anxiety.
University of Iowa College of Nursing, John A. Hartford Foundation Center of Geriatric Nursing Excellence | 9/25/2009
Mood, memory, and cognition. In: Menopause and osteoporosis update 2009.
Society of Obstetricians and Gynaecologists of Canada | 6/12/2009
  • CHEST Journal
    Print ISSN: 0012-3692
    Online ISSN: 1931-3543