Background
Over the last few decades, the increased intensity of treatments in patients with haematological malignancies (HMs) has translated into an improved survival rate. As a consequence of more aggressive chemotherapies and their complications, the need for patients with HMs for intensive care unit (ICU) support during the course of their disease has increased. Although outcomes of these patients have been fully investigated in terms of organ failure and mortality [
1], data are scarce on health-related quality of life (HRQOL) in this population [
2]. Long-term outcomes for physical and psychological factors, functional status and social interactions are becoming more and more important both for physicians and nurses and for patients and their relatives. Moreover, quality-of-life (QOL) measurements should be considered when making decisions on the allocation of healthcare resources.
In the ICU, HRQOL at admission has been shown to be inversely associated with multiple organ failure during hospitalization [
3] and hospital mortality [
4]. Moreover, baseline HRQOL has been shown to be correlated with HRQOL following discharge [
5,
6]. Finally, ICU survivors have significantly lower QOL compared with gender- and age-matched general population [
7]. Data on HRQOL in critically ill patients with HMs are scarce [
2]. Given the importance of the underlying disease as a strong predictor of HRQOL, the present study focused on critically ill haematological patients. We aim to assess post-ICU burden and HRQOL at 3 months and 1 year after ICU discharge and to identify risk factors for QOL impairment using the three widely used tools in this context: Short Form 36 (SF-36) [
8,
9], Hospital Anxiety and Depression Scale (HADS) [
10] and Impact of Event Scale (IES) [
11].
Discussion
This prospective study is the first multicenter study and the largest one focusing on HRQOL in critically ill patients with HMs.
To analyse all aspects of QOL, we used the SF-36 form, which offers a detailed mental and physical QOL evaluation and has been widely used and validated in ICU patients, the IES for post-traumatic stress syndrome evaluation and the HADS for evaluation of anxiety and depression.
In total, 1011 patients were included in the TRIALOH study. With a mortality rate of 50.7% (513 patients) at 3 months, 55.8% of survivors (278 patients) completed the SF-36 form at 3 months. Although response rates vary a lot between studies in ICU, ranging from 30 to 80% [
27,
28], our response rate is somewhat low (56% of the survivors) and may be explained in part by the high morbidity of patients with HMs, who require multiple hospitalizations and treatments in the course of their disease. Indeed, more days spent at hospital have been associated with no response at follow-up [
28]. Populations who are lost to follow-up or did not answer the questionnaire may have different demographic characteristics and different profiles of ICU-related morbidity than those who completed the QOL evaluation. Indeed, we found that age and the need for replacement therapy during ICU stay were 2 factors associated with the inability/refusal to complete the form. Different methods have been used during the study to improve response rates, such as sending hand written letters and employing reminders to the patients.
Compared with general ICU patients with septic shock, critically ill patients with HMs have profound alterations of HRQOL at 3 months. Moreover, although RP values improved at 1 year, global QOL impairment was consolidated at 1 year.
In a single-centre study, Oeyen et al. [
2] investigated long-term outcomes and QOL in critically ill patients with HMs (85 patients) or solid malignancies 3 months and 1 year after ICU discharge. Similarly, they found a profound alteration of QOL at 3 months and 1 year in HM patients with median PCS and MCS values even lower than those we observed in our study. Practices may have evolved over time, taking into account the QOL as an important goal when managing these patients. This may partly explain the discrepancy between the two studies.
In the present study, two main independent factors were significantly associated with QOL impairment (MCS) at 3 months: the SOFA score (i.e. organ failure) and the status of the underlying malignancy (complete or partial remission) at ICU admission. The initial SOFA score is used to quantify the degree of organ dysfunction present on admission, including the need for mechanical ventilation, vasopressors or renal replacement therapy. Many studies have demonstrated a strong correlation of initial SOFA score with mortality outcome [
15,
29,
30]. In our study, we have moved a step further, presenting SOFA score as a reliable predictor of QOL after ICU discharge. Combes et al. [
31] previously showed that prolonged invasive mechanical ventilation is associated with impaired HRQOL compared with that of a matched general population. Noninvasive strategies during ICU stay, such as decreasing the use of mechanical ventilation by using high flow oxygen, when appropriate, may improve QOL.
SOFA score has previously been associated with impaired QOL in critically ill patients with acute respiratory distress syndrome [
32]. In clinical practice, a crucial question is “for which critically ill HMs patients should we propose a limitation or withdrawal of life-sustaining treatments”? Decisions on withholding/withdrawing therapies should take into account the expected QOL after ICU discharge; the primary goal of intensive care being to return to a quality of life the patient would have found acceptable. Considering the SOFA score at admission as a predictor of both mortality and QOL, it may be helpful to physicians to inform patients and families in a reliable way and to guide in treatment decisions.
Paradoxically, disease status (partial or complete remission) at ICU admission was also associated with MCS impairment at 3 months. Patients in partial or complete remission have previously received intensive and potentially gruelling treatments for their underlying malignancy; conversely, most of patients classified as having an “evolutive disease” have newly diagnosed malignancies when admitted to the ICU. Indeed, patients who are considered in partial or complete remission had a longer time from the diagnosis of the underlying malignancy to ICU admission and, overall, have therefore received more chemotherapy than patients with evolutive disease.
The kinetic evaluation of QOL at 3 months and 1 year offers the opportunity to focus on aspects of QOL that may be improved by therapeutic interventions during the first year after ICU discharge. If some aspects of QOL (such as RP) have improved over time, we found that BP, GH, MH, RE scores and MCS decreased between 3 months and 1 year after ICU discharge. In general ICU populations, previous studies have shown that the majority of the QOL scores recover over time [
5]. Impairment of SF and RE suggests that patients with HMs may be socially isolated because of their condition. Association of a long hospital stay, including ICU stay and the presence of an underlying malignancy, has been shown to be major risk factors for social isolation [
33,
34].
During hospitalization, HM patients have to deal not only with treatment-related complications and adverse events, such as physical symptoms and changes in body image, but also with isolation-related psychological distress, including loss of control and lack of contact with family members and friends. Long-term psychological follow-up, social support and recreational activities during hospitalization may improve the QOL of these patients [
34,
35]. Centres that manage haematological patients may also consider extending visiting hours for relatives. Finally, bodily pain may be controlled by an approach carried out by a skilled pain care team and based on the association of causal therapies and adequate analgesics. PF, RP, VT and PCS are significantly lower in HMs patients compared with septic patients. Indeed, HMs and chemotherapies have been shown to induce a loss of weight and decrease in physical activity. Early mobilization in ICU and exercise may prevent the rapid loss of physical reserve and increase functional capacity [
36].
In the princeps study published in 2013 by Azoulay et al. [
1], 6 months after ICU discharge, the hematologists reported that all but seven ICU survivors were continuing their cancer treatment that ICU admission did not influence therapeutic intensity in 80% of ICU survivors, and that 80% of ICU survivors were in complete or partial remission. Unfortunately, precise data on post-ICU treatments are not available and we could not evaluate their impact on QOL.
Interestingly, only 8% of patients in our cohort were diagnosed with PTSD, according to the IES, which is lower than usually reported [
37]. The use of an ICU diary in many participating centres may have participated in the low incidence of PTSD as an ICU diary has been associated with a significant reduction in PTSD symptoms in critical illness survivors [
37]. A prospective multicenter comparative study of the impact of an ICU diary on PTSD in ICU is ongoing, involving many of the participating centres involved in this work.
The present study has several limitations. First, we were unable to provide QOL measures at ICU admission. Emergent admission and critical illness prevent patients from providing self-reported baseline QOL at the time of ICU admission. Reliable retrospective QOL data are difficult to obtain due to memory biases following ICU discharge. Answers to the question “Compared to 1 year ago, how would you rate your health in general now?” suggest that almost half of the patients felt better than 1 year before. Factors such as severity of current symptoms may bias patients’ recall of baseline status, especially in more severe patients, such as ICU patients [
38]. Indeed, Granja et al. [
39] have shown that almost 50% of ICU survivors did not remember the time in the hospital before ICU admission. Moreover, proxy assessment of patients’ baseline quality of life often differs from the patient’s assessment [
40]. Despite these limitations, HRQOL impairment at ICU admission has been correlated with multiple organ failure during ICU stay, increased hospital mortality and worsened HRQOL following discharge [
4,
6].
Second, as mentioned above, the response rate (55.8% of the survivors) is low, which is explained by the high morbidity of our population.
Third, our data are now 7 years old. Improvements in outcomes of cancer patients, as well as the improvements in process of care may have modified the impact of ICU complications on HRQOL. In HM patients, several cohort studies and trials evaluating a noninvasive diagnostic and therapeutic management of acute respiratory failure have shown the feasibility and safety of the noninvasive strategy [
41]. High flow oxygen has also demonstrated survival benefits as compared to noninvasive ventilation [
42]. The increased use of noninvasive strategies may then have improved HRQOL of these patients after ICU discharge.
Finally, a longer follow-up would have been interesting to analyse, as a gradual improvement of most aspects of QOL might have occurred after 1 year.
Authors’ contributions
FE, LZ and EA contributed to conception and design. FE, LZ, EA, VL, DM, FP, AK, JM, FV, MN, FB, AR, CL, PP, APM, DB, RH, MD, DC and NDP were involved in provision of study materials or patients, collection and assembly of data. LB, FE, LZ and EA were involved in data analysis and interpretation. FE, LZ and LB contributed to manuscript writing. All authors read and approved the final manuscript.