Quality of life in survivors after a period of hospitalization in the intensive care unit: a systematic review

Objective To assess the long-term, health-related quality of life of intensive care unit survivors by systematic review. Methods The search for, and selection and analysis of, observational studies that assessed the health-related quality of life of intensive care unit survivors in the electronic databases LILACS and MEDLINE® (accessed through PubMed) was performed using the indexed MESH terms "quality of life [MeSH Terms]" AND "critically illness [MeSH Terms]". Studies on adult patients without specific prior diseases published in English in the last 5 years were included in this systematic review. The citations were independently selected by three reviewers. Data were standardly and independently retrieved by two reviewers, and the quality of the studies was assessed using the Newcastle-Ottawa scale. Results In total, 19 observational cohort and 2 case-control studies of 57,712 critically ill patients were included. The follow-up time of the studies ranged from 6 months to 6 years, and most studies had a 6-month or 1-year follow up. The health-related quality of life was assessed using two generic tools, the EuroQol and the Short Form Health Survey. The overall quality of the studies was low. Conclusions Long-term, health-related quality of life is compromised among intensive care unit survivors compared with the corresponding general population. However, it is not significantly affected by the occurrence of sepsis, delirium, and acute kidney injury during intensive care unit admission when compared with that of critically ill patient control groups. High-quality studies are necessary to quantify the health-related quality of life among intensive care unit survivors.


INTRODUCTION
Technological advances in the intensive care have increasingly reduced intensive care unit (ICU) mortality. (1) However, the consequences of a critical illness can persist for a long time, affecting the physical, cognitive and mental health of ICU survivors. (2) The multiplicity of these consequences was recognized as "postintensive care syndrome" and may have a strong, negative impact on functioning and on health-related quality of life (HRQOL). (2) Assessing outcomes related to physical and psychological factors, functional status, social interaction and HRQOL is as important as assessing the long-term mortality rate of ICU survivors. (3,4) Ideally, ICU survivors should reach their premorbidity and/ or admission HR-QOL scores or even reach scores that are better than or similar to those of age-, sex-and comorbidity-matched individuals. (5) Although HRQOL scores are increasingly included in studies and recognized as important outcome parameters in this population, such results generate inaccuracy in their interpretation for various reasons. First, the period during which HRQOL recovery should occur in not defined, and therefore, the optimal follow-up period for HRQOL evaluation remains undetermined. Frequently, postintensive care syndrome may clinically manifest as not only transient events, (6,7) occurring 5 years after hospital discharge, (8) but also permanent events, at least for some survivors. Second, baseline HRQOL evaluation is difficult, thus complicating critical illness burden investigations. In addition, the evaluation instruments used in the studies are different, thereby complicating the systematization and interpretation of HRQOL results. However, a better understanding of how much intensive care affects the long-term HRQOL of ICU survivors is necessary to help healthcare professionals in making decisions on future efforts to reduce the burden of critical illness.
The objective of this study was to perform a systematic review, evaluation and synthesis of observational studies published in the literature on the long-term HRQOL of ICU survivors in comparison with the corresponding general population and control groups.

METHODS
This systematic review was performed according to the Preferred Reporting Items for Systematic Reviews and Meta--Analyses Guidelines (PRISMA). (9) A systematic search strategy with the indices "quality of life [MeSH Terms]" AND "critically illness [MeSH Terms]" was used in the electronic databases, Latin American and Caribbean Health Sciences Literature (Literatura Latino-Americana e do caribe em Ciências da Saúde -LILACS) and Medical Literature Analysis and Retrieval System Online (MEDLINE®), which were accessed through PubMed from October 13, 2016, to November 7, 2016.
The titles and abstracts of the articles identified in the search strategy were evaluated based on eligibility criteria (Table 1) by three independent reviewers. Full-text articles were downloaded when the abstract lacked information on inclusion and exclusion criteria. The list of references of the selected articles and the personal files of the researchers were also searched for to identify possible studies that might also meet the eligibility criteria of the study and that might not have been found in the initial search. Any discrepancies between the reviewers were resolved by consensus, and a fourth reviewer assessed the publications in cases of persistent disagreement.
Data were independently retrieved from the selected articles by two reviewers. Discrepancies were resolved by consensus or by a third reviewer.
The methodological quality of the studies was evaluated using the Newcastle-Ottawa scale by two independent, previously trained and qualified reviewers. The methodological quality score of the cohort and case-control studies was calculated based on three components: study group selection (0 -4 points), quality of adjustment for confounding factors (0 -2 points) and evaluation of exposure or outcome of interest (0 -3 points). The maximum score was 9 points, which expressed high methodological quality. Discrepancies between the reviewers were resolved by consensus, and another evaluation was performed by a third reviewer in case of persistent disagreement.

Study selection
The database searches identified 417 references, and the consultation of other sources identified 5 references.
Of these references, 289 were excluded because they were published outside of the 5-year period that was stipulated for this review. Of the other references, 112 were excluded after reading the title and the abstract, and eventually the full text, because they failed to meet the other inclusion criteria (Figure 1). There was no discrepancy in the number of articles selected by the three reviewers, and 21 articles were included.
In 18 of the 21 studies included in this review, the long-term HRQOL of ICU survivors was compromised when compared with that of the corresponding general population. In studies with a 6-month follow-up time, the HRQOL of critically ill and elderly patients with low severity scores and critically ill patients with acute kidney injury was similar to that of the corresponding general population. (10,12,26) Most HRQOL dimensions improved in the long run. (10,16,18,19,22,24,26,29) The domains related to physical aspects were the most affected. (13,18,19,22,23,27)

Risk of bias in the included studies
The general methodological quality of the studies included in this review was low ( Table 3). The Newcastle--Ottawa scores of the studies ranged from 2 to 5; a score lower than 4 indicated limited or low-quality evidence. Consensus was reached on all occasions, and no study was excluded from this review, based on the risk of bias assessed. Meta-analyses could not be performed because the studies included in this review had a predominantly observational cohort design.

DISCUSSION
This systematic review describes the long-term HR-QOL of ICU survivors. In total, 21 studies were included in this review. The overall quality of the studies was low, according to the Newcastle-Ottawa scale, thus highlighting the need for studies with high methodological quality to determine the long-term HRQOL of ICU survivors. New studies with appropriate methodological designs may provide important data on the main factors that result in their change, as well as on possible therapeutic alternatives.

Critically ill patients
The long-term HRQOL of the critically ill patients differed among the studies analyzed, varying by population and follow-up time. In the short term, the mental component of the HRQOL in a population of critically ill patients with a low severity score was similar to that of the corresponding age-matched population at 8 and 26 weeks after hospital discharge. (26) In a 1-year follow--up, three studies, which were conducted in Holland, (17) Argentina (29) and the United States, (25) highlighted that HRQOL was significantly compromised and that it was even more affected among specific subgroups of ICU survivors, such as those diagnosed with shock who remained under mechanical ventilation for a long period and who showed persistent weakness. (17,29) Conversely, a study that compared the impact on HRQOL between hospital and ICU patients highlighted that HRQOL is clinically impaired in both groups 1 year after discharge, with no significant difference between hospital and ICU patients. (25) In the 5-year period, after correcting for natural decline, the HRQOL of ICU survivors significantly decreased, and the physical and social functioning and overall health domains of ICU survivors remained significantly lower than those of the age-matched general population. (18) However, the size effect of the HRQOL reduction was weak on all domains of the evaluation instrument, thus suggesting that the ICU admission effect on the perception of HRQOL 5 years after discharge may not be clinically relevant.  (22) Urinary: 3 (6)

Elderly critically ill patients
Among elderly ICU survivors, the HRQOL 1 year after ICU discharge was worse than that of the age- (13) and age-and sex-matched (23) general population. Conversely, in another study (10) with a shorter follow-up time, which was performed in an elderly population with lower severity scores, the HRQOL scores 6 months after ICU discharge were similar both to the scores before ICU admission, which were obtained by a patient representative, and to those of the age-matched general population. In both studies (13,23) that showed impaired HRQOL, physical functioning was the most affected domain. Understandably, elderly ICU survivors show impaired HRQOL given their likely increase in comorbidities, and this impairment was more visible in physical functioning.

Critically ill patients with sepsis
The HRQOL of critically ill patients with sepsis was not significantly different from that of nonelderly critically ill patients with sepsis, (30) of critically ill patients with community-acquired sepsis (14) or of age-, sex-and Charlson comorbidity index-matched critically ill patients without sepsis, (28) both in 6-month (14) and in 1- (30) and 6-year follow--up studies. (28) However, the HRQOL of these patients showed a clinically relevant impairment when compared with the general population. (14,(19)(20)(21)28) Approximately 50 to 75% patients with sepsis (7) progressed with ICU-acquired muscle weakness, which is one of the main signs of physical function impairment of postintensive care syndrome, versus 25 to 50% patients subjected to mechanical ventilation. (7) However, the short- (14) and long-term (28,30) HRQOL of these patients was not significantly different in comparison with critically ill patients with other diagnoses, thus showing that ICU admission, regardless of diagnosis and patient clinical status, is the determinant of impaired HRQOL in these survivors.
cognitive problems. (11) Previous studies (31,32) have shown that delirium during an ICU stay is associated with long--term cognitive deficit and mortality, leading to speculation that delirium would also affect the long-term HR-QOL, which has not been fully elucidated yet.

Critically ill patients with acute kidney injury treated with renal replacement therapy
Two studies (12,16) investigated the HRQOL of critically ill patients with acute kidney injury treated with renal replacement therapy. Both studies found no long-term HR-QOL differences between critically ill patients with acute kidney injury and those without. However, the studies differed when comparing the HRQOL of critically ill patients with acute kidney injury with the HRQOL of the healthy population. Vaara et al. (12) found no differences at 6 months, whereas Oeyen et al. (16) found significant differences after 4 years of follow-up. Importantly, Vaara et al. (12) conducted a retrospective short-term study.

Critically ill patients with acute lung injury
Only one study (22) evaluated the HRQOL of critically ill patients with acute lung injury. In this population, which was evaluated in the United States, the baseline values of the physical functioning component of the HRQOL were substantially lower than those estimated at 2 years of follow-up. A previous and highly relevant study (8) on the subject demonstrated that ICU survivors who developed acute respiratory distress syndrome showed physical and psychological sequelae and a reduced physical function component of the HRQOL 5 years after ICU discharge, corroborating the finding of the study included in this review.

Critically ill patients with poor sleep quality
In individuals with poor sleep quality, the only study published on HRQOL in this population demonstrated that the physical and mental functioning components of the evaluation instrument were significantly lower in these individuals 6 months after ICU discharge. (27) Evidence indicates that low quality of ICU sleep and acute sleep deprivation lead to possible negative effects on recovery in critically ill patients, (33,34) including physical and psychological recovery.

CRITICALLY ILL PATIENTS WITH FRAILTY
Frailty is a multidimensional state characterized by physiological and cognitive loss in older patients, and it predicts adverse events and unfavorable outcomes. (35) Critically ill patients in a state of frailty classified with the Canadian Study of Health and Ageing Clinical Frailty Scale were evaluated in a multicenter cohort in Canada 1 year after ICU admission. (24) These individuals showed worse HRQOL scores than nonfrail individuals and the healthy population. (24) Another study included in this review, performed with elderly people aged 80 years or older, demonstrated that frailty was an independent predictor that was more significant than age, critical illness severity or comorbidities -which are commonly considered key determinants of long-term outcome. (23) Postintensive care syndrome A substantial, albeit unknown, proportion of ICU survivors is at risk of developing postintensive care syndrome. Increasing efforts have been made to use the term "postintensive care syndrome" to describe new or aggravated physical, cognitive or mental deficits, resulting from critical illness, that persist after hospitalization, (2,36) with the aim of understanding the epidemiology of this syndrome and its burden on the long-term HRQOL of ICU survivors. Approximately 25 to 50% of patients subjected to mechanical ventilation will develop ICU-acquired muscle weakness, (7) and approximately 85 to 95% of them persist with neuromuscular deficits for 2 to 5 years after hospital discharge. (7) Approximately 30 to 80% patients show cognitive impairment after their ICU stay, (37) and 10 to 50% patients show new depression, anxiety, posttraumatic stress and sleep disorder symptoms. (7) The high and persistent prevalence of changes related to postintensive care syndrome apparently justifies the long-term negative effects on the HRQOL of ICU survivors, and these consequences are more prominent in some specific situations found in the intensive care setting, as shown in this systematic review.
Importantly, the long-term HRQOL assessment presented by these studies clearly disregarded patients who dropped out or died, as only 42% (24,200 patients) of the total sample (57,712) was assessed. We believe that this finding compromises the HRQOL assessment of this population because sample loss may be related to worsened patient clinical status or death.

RESUMO
Descritores: Qualidade de vida; Estado terminal; Cuidados críticos; Unidades de terapia intensiva; Tempo de internação The search strategy used in electronic databases failed to identify some eligible studies. Previous systematic reviews (38,39) on the subject used broader search strategies such as the following: ("quality of life" OR "health status indicators") AND ("intensive care units" OR "critical care" OR "critical illness" OR "sepsis" OR "adult respiratory distress syndrome") and ("quality of life" OR "long-term outcome") AND ("intensive care" OR "critical care" OR "critically ill patients" OR "ICU patients" OR "critical care patients" OR "ICU stay" OR "ICU"). Dowdy et al. (38) included the terms "sepsis" and "acute respiratory distress syndrome" in the search strategy because an eligible study identified before conducting the search was not identified when using the initial terms. However, we complemented the search in the reference lists of other systematic reviews and relevant publications on the subject.
The scope of this review was comprehensive; therefore, the heterogeneity of the studies was a limitation, precluding their comparison. We chose to broaden our review to better describe the potential burden of ICU hospitalization on long-term HRQOL. Future, high-quality studies in specific populations are necessary to prepare meta-analyses for specific ICU groups.

FINAL CONSIDERATIONS
Long-term, health-related quality of life is compromised among intensive care unit survivors when compared with the corresponding general population. However, long-term, health-related quality of life is not significantly affected by the presence of sepsis, delirium or acute kidney injury during intensive care unit stay when compared with that of critically ill patient control groups. High-quality studies are necessary to quantify the health-related quality of life of intensive care unit survivors.