Abstract
Background
Mortality of immunocompromised patients is particularly high in intensive care units (ICUs) and mainly depends on severity at admission. Moreover, mortality is also high during the months following ICU discharge. The reasons for these poor outcomes after ICU discharge have not been adequately studied.
Research question
We hypothesized that the factors associated with poor outcomes after ICU discharge of immunocompromised patients would be different from those associated with in-ICU mortality.
Study design and methods
This is a post-hoc analysis of a multicenter clinical trial comparing two noninvasive oxygenation strategies in immunocompromised patients admitted to ICU for acute hypoxemic respiratory failure. Multivariable analyses were performed to determine early factors (i.e within 6 h of admission) associated with in-ICU mortality, as well as factors associated with poor functional outcomes (i.e death or survival with poor performance status) at 6 months, only in ICU survivors.
Results
Among the 299 patients analyzed, the mortality rate was 31% (94 patients) in the ICU and 49% at 6 months (146 patients). Solid cancer (adjusted odds ratio 2.92 [95% confidence interval, 1.22–7.28]), severity SOFA score at admission (aOR 1.29 [1.14–1.48]), the extent of pulmonary infiltrates on chest X-ray (aOR 1.57 [1.17–2.15]) and increased discomfort one hour after initiation of noninvasive respiratory support (aOR 2.08 [1.12–3.85]) were independently associated with in-ICU mortality. Out of the 202 ICU survivors whose performance status was reported, solid cancer (aOR 3.03 [1.33–9.09]) and poor performance status before ICU admission (aOR 2.43 [1.03–5.88]) were both associated with poor outcome at 6 months, independently from the decision to forgo life-sustaining therapies (aOR 5.88 [2.17–20.00]).
Interpretation
Whereas in-ICU mortality of immunocompromised patients with acute respiratory failure was mainly driven by severity, poor outcomes at 6 months were mainly driven by performance status before ICU admission. Solid cancer was independently associated with both poor short as well as longer-term outcomes.
Trial registration Clinical trial registration: NCT04227639
Supplementary Information
The online version contains supplementary material available at 10.1186/s13613-025-01578-1.
Keywords: Immunosuppression, Acute respiratory failure, Intensive care unit, Mortality
Introduction
Immunosuppression is defined as a dysfunction of the immune system, whether primary or secondary, resulting from a medical condition such as solid cancer, hematologic malignancy, human immunodeficiency virus, or immunosuppressive drugs [1]. The prevalence of immunosuppression has been steadily increasing in the recent years, due primarily to the rising incidence of cancer and the improved survival of these patients, stemming from earlier detection of diseases and therapeutic advances [2]. Therefore, it can be anticipated that a considerable proportion of patients will live several years with various types and degrees of immunosuppression, exposing them to an increased risk of developing severe life-threatening infections [3]. As a result, immunocompromised patients account for 15% to 25% of all patients admitted to intensive care units (ICUs) [4].
Acute respiratory failure is the main reason for ICU admission in this subset of patients [5]. Their mortality in the ICU is high, approximating 30% in overall [6], and may exceed 50% in those requiring invasive mechanical ventilation [7]. In addition, their mortality after ICU discharge continues to increase, reaching about 50% at six months and over 60% at one year after ICU admission [8, 9]. However, risk factors for mortality among ICU survivors have not been adequately studied. Indeed, most of the studies assessing factors associated with long-term mortality in immunocompromised patients admitted to the ICU have to a large extend based their analyses on factors associated with ICU mortality, such as the severity of organ failure [8, 9]. Additionally, although they are of major importance when selecting an adequate therapeutic option in cancer or hematology patients [8, 10, 11], the factors influencing functional status after discharge have only rarely been assessed in immunocompromised patients admitted to the ICU for acute hypoxemic respiratory failure [12]. Accounting for the functional status assessment of immunocompromised patients after ICU discharge would be of primary importance in order to better understand which patients would benefit from ICU admission or readmission.
We hypothesized that among immunocompromised patients admitted to an ICU for acute respiratory failure, the factors associated with ICU mortality outcomes differed from those associated with poor outcomes after ICU discharge.
Methods
Study design
We performed an unplanned post-hoc analysis of a randomized controlled trial conducted in 29 centers in France and in Italy comparing two noninvasive oxygenation strategies in immunocompromised patients admitted to the ICU for acute hypoxemic respiratory failure [7]. The protocol was approved by the Ethics Committee Ouest III (Poitiers, France) for French centers and by the local ethics committee for the Italian center. Informed consent from patients or their surrogate was obtained orally, with a written record maintained by the investigator.
Patients
Adult (≥ 18 years old) immunocompromised patients admitted to the ICU for acute hypoxemic respiratory failure defined as a respiratory rate ≥ 25 breaths per minute and PaO2/FiO2 ratio ≤ 300 mmHg while spontaneously breathing with standard oxygen (oxygen flow rate ≥ 10 L/min), with high-flow nasal oxygen therapy or with noninvasive ventilation were included. Immunosuppression was defined by one of the following criteria: hematological malignancy (active or remitting < 5 years), allogenic stem cell transplantation within the previous 5 years, active or relapsing solid cancer, leucopenia < 1 G/L or neutropenia ≤ 0,5 /induced by chemotherapy, solid organ transplantation, acquired immunodeficiency syndrome, systemic steroids ≥ 0.5 mg/kg per day of prednisone equivalent for at least 3 weeks, or immunosuppressive or immunomodulatory drugs. The main exclusion criteria were patients who could strongly benefit from noninvasive ventilation (hypercapnia defined as PaCO2 > 50 mmHg-, acute-on-chronic respiratory failure, cardiogenic pulmonary edema, surgery less than 7 days before), shock, impaired consciousness, urgent need for intubation, and Do Not Intubate order at time of inclusion. Additionally, for the current analysis, we excluded patients in whom Eastern Cooperative Oncology Group (ECOG) performance status at 6 months after randomization was missing.
Classification of immunosuppression subgroups
Patients were classified according to the underlying cause of immunosuppression in 3 subgroups: (1) active or remitting for less than 5 years hematological malignancy whatever the treatment (chemotherapy alone, autologous stem cell transplantation, allogeneic stem cell transplantation), (2) active solid cancer and (3) other causes (i.e. acquired immune deficiency syndrome, solid organ transplant recipient or immunosuppressive drugs for connective tissue disease). For patients with more than one cause of immunosuppression (for instance a solid organ transplant recipient who developed lymphoma), the most recent disease that led to ICU admission for acute hypoxemic respiratory failure was considered as the main cause of immunosuppression.
Data collection
Demographic characteristics (age, sex, comorbidities, cause of immunosuppression, ECOG performance status before ICU admission), cause of acute respiratory failure, severity scores including Simplified Acute Physiology Score (SAPS) II at ICU admission [13] and Sequential Organ Failure Assessment (SOFA) at admission [14] clinical characteristics and discomfort score (by means of a 100 mm visual analogue scale ranging from 0 - “no discomfort” – to 100 – “maximal imaginable discomfort”) before treatment initiation and after one hour of treatment, presence of bilateral infiltrates and the number of quadrants with infiltrates on chest X- ray at inclusion according to the investigator, treatments during ICU stay (need for intubation, dialysis, or chemotherapy, and decision to forgo life-sustaining therapies) and outcomes (mortality at day 28, at 6 months, and ECOG performance status at 6 months) were prospectively collected.
Study outcomes
Poor outcome at 6 months was defined as death or survival with ECOG performance status of 3 or 4 at 6 months after admission among ICU survivors.
Statistical analysis
Continuous variables were expressed as mean ± standard deviation (SD) or median (25–75th percentiles) depending on their distribution and compared using the Student t-test or the Mann-Whitney U test, as appropriate. Categorical variables were expressed as number (percentage) and compared using the Chi-squared test or Fisher test, as appropriate. A first stepwise backwards logistic regression model was computed to identify early variables (i.e. occurring within the first 6 h after admission) associated with ICU mortality in the overall population. Intubation within the first 24 h after ICU admission and SAPS II were not included in the model despite their strong association with ICU mortality [6, 13] because these variables require at least 24 h after ICU admission to be considered, hampering their clinical utility as early markers. A second stepwise backward logistic regression model was computed to identify the variables associated with poor outcomes at 6 months in the subset of ICU survivors. Variables included in the multivariable analysis included noncollinear variables associated with poor outcomes with a p value < 0.20 in univariate analysis. Variables known to be associated with poor outcomes and the noninvasive oxygenation strategy allocated were also forced in the models. Statistical analyses were conducted by using R software version 4.3.2 (R Foundation for Statistical Computing; https://www.R-project.org). Two-tailed p values < 0.05 were considered significant.
Results
Among the 299 immunocompromised patients included in the seminal study, malignancy (either hematological or solid) was the cause of immunosuppression in 74% of cases (222 patients). Overall, ECOG performance status before ICU admission was 1 ± 1 in mean, including 16% of patients (49 out of 299) with a poor score of 3 or 4. At admission, severity SAPS-II and SOFA scores were 45 ± 16 and 6 ± 3, respectively. At inclusion, the respiratory rate was 31 ± 5 breaths/min, PaO2/FiO2 ratio 147 ± 56 mmHg, and 74% of patients (223 out of 299) had bilateral infiltrates on chest X-ray. Characteristics of patients according to the immunosuppression subgroup is displayed in Additional Table 1.
Mortality in ICU and at 6 months was 31% (94 patients) and 49% (146 patients), respectively. Among the 205 ICU survivors, and after excluding three patients whose ECOG performance status was missing at 6 months (Additional Table 2.), 31% had a poor outcome at 6 months (62 out of 202 patients), including 52 patients who died after ICU discharge (26%) and 10 patients who were alive with ECOG performance status 3 or 4 (5%) at 6 months (Fig. 1).
Fig. 1.
Flow chart of patients included in the analysis
Factors associated with ICU mortality
Using univariate analysis, patients who died in the ICU were more likely to have higher severity scores, a higher number of quadrants with infiltrates on chest X-ray, higher respiratory rate and more severe hypoxemia at admission. Additionally, they were more likely to have an increased in discomfort score after treatment initiation than patients who survived in the ICU (Table 1 and Fig. 2).
Table 1.
Univariate analysis of factors associated with ICU mortality
| Variables | Death in the ICU (n = 94) |
Alive at ICU discharge (n = 205) |
p value |
|---|---|---|---|
| Baseline characteristics at admission | |||
| Age, years | 65 ± 12 | 63 ± 12 | 0.111 |
| Male sex, n (%) | 62 (66%) | 130 (63%) | 0.767 |
| Body mass index, kg/m2 | 25 ± 5.2 | 25 ± 5.8 | 0.670 |
| Body mass index < 18.5 kg/m2, n (%) | 6 (6.4%) | 19 (9.3%) | 0.566 |
| Simplified Acute Physiology score II | 51 ± 20 | 43 ± 14 | < 0.001 |
| ECOG performance status 3 or 4, n (%) | 20 (21%) | 29 (14%) | 0.168 |
| Charlson comorbidity score | 3.6 ± 2.6 | 3.4 ± 2.3 | 0.585 |
| Underlying condition, n (%) | |||
| Type of immunosuppression | 0.085 | ||
| Hematological malignancy | 47 (50%) | 103 (50%) | |
| Solid cancer | 29 (31%) | 43 (21%) | |
| Other cause A | 18 (19%) | 59 (29%) | |
| Characteristics at inclusion | |||
| SOFA score | 6.8 ± 3.0 | 5.7 ± 2.5 | 0.002 |
| SOFA score without respiratory item | 3.6 ± 3.0 | 2.7 ± 2.4 | 0.012 |
| Need for norepinephrine, n (%) | 10 (11%) | 8 (3.9%) | 0.044 |
| Thrombocytopenia, n (%) | 12 (13%) | 32 (16%) | 0.639 |
| Respiratory rate, breaths/min | 32 ± 5 | 31 ± 5 | 0.072 |
| pH, units | 7.43 ± 0.09 | 7.44 ± 0.06 | 0.406 |
| PaO2/FiO2, mmHg | 134 ± 51 | 154 ± 58 | 0.003 |
| PaCO2, mmHg | 34 ± 6 | 34 ± 6 | 0.388 |
| Discomfort score, mm B | 46 ± 27 | 45 ± 28 | 0.892 |
| Bilateral infiltrates on chest X-ray, n (%) | 75 (80%) | 148 (72%) | 0.209 |
| Number of quadrants with infiltrates on chest X-ray | 3.1 ± 1 | 2.8 ± 1.1 | 0.014 |
| 1 h after treatment initiation | |||
| Randomization in the noninvasive ventilation arm, n (%) | 49 (52%) | 96 (47%) | 0.468 |
| Respiratory rate, breaths/min | 30 ± 8 | 27 ± 7 | 0.023 |
| pH, units | 7.42 ± 0.10 | 7.45 ± 0.06 | 0.067 |
| PaO2/FiO2, mmHg | 146 ± 73 | 183 ± 92 | 0.001 |
| PaCO2, mmHg | 35 ± 6 | 34 ± 6 | 0.640 |
| Discomfort score, mmC | 45 ± 29 | 39 ± 27 | 0.129 |
| Change between inclusion and H1, n (%) | |||
| Increased respiratory rate | 37 (39%) | 70 (34%) | 0.485 |
| Increased PaO2/FiO2 | 41 (44%) | 108 (53%) | 0.181 |
| Increased PaCO2 | 51 (54%) | 98 (48%) | 0.181 |
| Increased discomfort scoreC | 46 (49%) | 69 (34%) | 0.045 |
Qualitative variables are expressed in number (percentage), quantitative variables are expressed in mean ± standard deviation or median [25th−75thpercentile] according to their distribution.
ECOG: Eastern Cooperative Oncology Group; SOFA: Sequential organ failure assessment; PaO2: partial pressure of arterial oxygen; FiO2: fraction of inspired oxygen, PaCO2: partial pressure of arterial carbon dioxide; H1: 1 h after randomization.
A Acquired immune deficiency syndrome, solid organ transplant or immunosuppressive drugs for connective tissue disease.
Discomfort score was assessed using a 100 mm visual analogue scale ranging from no discomfort (0) to maximum imaginable discomfort (100).
C H1 refers to one h after randomization.
D assessed in 250 patients.
Fig. 2.
Alluvial plot of the trajectories of immunocompromised patients admitted to the ICU for acute respiratory failure according to their subtype of immunosuppression, up to ICU discharge and 6 months
Using multivariable analysis, solid cancer, radiographic severity defined by the number of quadrants with pulmonary infiltrates on chest X-ray and increased discomfort score after treatment initiation were associated with ICU mortality, independently from severity indicated by SOFA score (Fig. 3A).
Fig. 3.
Forrest plot of factors independently associated with ICU mortality (panel A) and poor outcome at 6 months among ICU survivors (Panel B). Panel A: Variables included in the backward stepwise logistic regression model were age, ECOG performance status 3 or 4 before ICU admission, solid cancer, SOFA score at inclusion, number of quadrants with infiltrates on chest X-ray, respiratory rate at inclusion, respiratory rate and pH 1 h after treatment initiation, and increased PaO2/FiO2, PaCO2 and discomfort score between inclusion and 1 h after treatment initiation. The noninvasive oxygenation strategy allocated in the seminal study was forced in the model and interaction between variables was tested. (Hosmer-Lemeshow goodness-of-fit test: p = 0.41) Panel B: Variables included in the backward stepwise logistic regression model were ECOG performance status 3 or 4 before ICU admission, body mass index < 18.5 kg/m2, solid cancer and patients with other causes of immunosuppression, microbiologically documented lung infection, no etiological diagnosis of respiratory failure, decision to forgo life-sustaining therapies during ICU stay, and ICU length of stay in the ICU. Intubation was forced in the model. The noninvasive oxygenation strategy allocated in the seminal study was forced in the model and interaction between variables was tested. (Hosmer-Lemeshow goodness-of-fit test: p = 0.86)
Factors associated with poor outcome at 6 months among ICU survivors
Using univariate analysis, patients with poor outcome after ICU discharge (i.e. who died or who were alive with an ECOG performance status of 3 or 4 at 6 months) were more likely to have solid cancer or a cause of immunosuppression other than hematologic disease. More patients with a poor outcome at 6 months had already ECOG performance status of 3 or 4 before ICU admission, decision to forgo life-sustaining therapies during ICU stay and to have no etiological diagnosis of respiratory failure than patients with good outcome at 6 months (Table 2 and Fig. 2).
Table 2.
Univariate analysis of factors associated with poor outcome at 6 months among ICU survivors
| Variables | Died or alive with ECOG 3 or 4 at 6 months (n = 62) |
Alive with ECOG ≤ 2 at 6 months (n = 140) |
p value |
|---|---|---|---|
| Baseline patient characteristics | |||
| Age, years | 62 ± 12 | 63 ± 13 | 0.614 |
| Sex, male, n (%) | 37 (60%) | 90 (64%) | 0.640 |
| Body mass index, kg/m2 | 24 ± 5.8 | 25 ± 5.8 | 0.211 |
| Body mass index < 18.5 kg/m2, n (%) | 10 (16%) | 9 (6%) | 0.058 |
| Simplified Acute Physiology score II | 43 ± 14 | 42 ± 14 | 0.621 |
| ECOG performance status 3 or 4, n (%) | 14 (23%) | 15 (11%) | 0.045 |
| Charlson comorbidity score | 3.8 ± 2.7 | 3.2 ± 2.0 | 0.088 |
| Underlying condition, n (%) | |||
| Type of immunosuppression | < 0.001 | ||
| Hematological malignancy | 28 (45%) | 74 (53%) | |
| Solid cancer | 23 (37%) | 19 (14%) | |
| Other a | 11 (18%) | 47 (34%) | |
| During intensive care unit stay | |||
| Randomization in the noninvasive ventilation arm, n (%) | 28 (45%) | 67 (48%) | 0.840 |
| ARDS according to the new definition at randomization, n (%) | 0.999 | ||
| No or mild ARDS | 26 (42%) | 58 (41%) | |
| Moderate or severe ARDS | 36 (58%) | 82 (59%) | |
| Bronchoalveolar lavage, n (%) | 20 (32%) | 57 (41%) | 0.670 |
| CT-scan, n (%) | 37 (60%) | 91 (65%) | 0.572 |
| Change in immunosuppressive drug regimen, n (%) | 18 (29%) | 36 (26%) | 0.750 |
| Renal replacement therapy, n (%) | 7 (11%) | 12 (8.6%) | 0.727 |
| Need for norepinephrine within 3 days after randomization, n (%) | 8 (13%) | 11 (7.9%) | 0.365 |
| Intubation, n (%) | 16 (26%) | 43 (31%) | 0.589 |
| Duration of mechanical ventilation, days | 10 ± 7 | 12 ± 9 | 0.433 |
| Reintubation, n (%) | 1 (6%) | 3 (7%) | 1.000 |
| Decision to forgo life-sustaining therapies, n (%) | 16 (26%) | 6 (4.3%) | < 0.001 |
| Cause of respiratory failure, n (%) | |||
| Microbiologically documented lung infection | 24 (39%) | 79 (56%) | 0.030 |
| Specific | 11 (18%) | 13 (9.3%) | 0.140 |
| Toxic cause | 3 (4.8%) | 8 (5.7%) | 0.999 |
| Cardiogenic pulmonary edema | 2 (3.23%) | 9 (6.4%) | 0.556 |
| Miscellaneous | 5 (8.1%) | 12 (8.6%) | 0.999 |
| No diagnosis | 15 (24%) | 18 (13%) | 0.071 |
| Short-term outcomes | |||
| Length of stay in the ICU, days | 7 [5–12] | 8 [5–14] | 0.122 |
| Length of stay in the hospital, days | 22 [14–36] | 19 [13–31] | 0.596 |
| Length of hospital stay after ICU discharge, days | 12 [5–22] | 10 [6–18] | 0.542 |
Qualitative variables are expressed in number (percentage), quantitative variables are expressed in mean ± standard deviation or median [25th−75th percentile] according to their distribution
ECOG: Eastern Cooperative Oncology Group; ARDS: acute respiratory distress syndrome; CT-scan: computed tomography scan; ICU: intensive care unit
a Acquired immune deficiency syndrome, solid organ transplant or immunosuppressive drugs for connective tissue disease
Using multivariable analysis, solid cancer and ECOG performance status 3 or 4 before ICU admission were associated with poor outcome at 6 months, independently from the decision to forgo life-sustaining therapies (Fig. 3B).
Discussion
In this post-hoc analysis of a randomized trial including immunocompromised patients admitted to an ICU for acute hypoxemic respiratory failure, severity SOFA score at admission, solid cancer, radiographic severity and increased discomfort after treatment initiation were independently associated with ICU mortality. Among admitted patients, ICU mortality was 31%. Of those discharged from the ICU, 31% either died before 6 months or had poor performance status at that time point.
After ICU discharge, solid cancer, ECOG performance status 3 or 4 before ICU admission and decision to forgo life-sustaining therapies during ICU stay were the three variables independently associated with poor outcomes at 6 months.
In our study, in-ICU mortality was 31%, in keeping with the 32% in-ICU mortality rate reported by in a large-scale prospective observational international cohort study including 1611 immunocompromised patients similar to ours (52% of patients with hematological malignancy and 35% with solid cancer) [6]. Overall mortality rate at 6 months was 49% which is also close to the 47% mortality rate reported in a retrospective study including 366 patients with hematological malignancy [15], and to the 55% mortality rate reported in a multicenter observational study including 449 patients with lung cancer [8]. All in all, these findings reinforce the external validity of our findings.
In addition to the severity score assessed at admission, we found that solid cancer, radiographic severity and increased discomfort after treatment initiation were three independent predictors of death in the ICU. Severity of organ failure has been extensively reported as a risk factor for ICU mortality in patients with hematological malignancy [16], lung cancer [8, 17], and in a mixed population of immunocompromised patients admitted to an ICU for acute respiratory failure [6]. Although scarcely compared in the literature, mortality of patients with solid cancer has been shown to be higher than that of patients with hematological malignancy [18, 19]. Unfortunately, due to the lack of power of our study, we were not able to compare outcomes according to types of solid cancer, a major determinant of short-term prognosis [20]. Likewise, the number of quadrants with pulmonary infiltrates on chest X-ray, indicating the radiologic severity of the respiratory disease, has previously been associated with increased risk of intubation or death in immunocompromised patients with acute respiratory failure [21, 22]. Interestingly, increased discomfort score one hour after initiation of noninvasive oxygenation strategy, a simple tool to use at the bedside, was independently associated with ICU mortality. This discomfort score may, at least partially, reflect dyspnea, an independent risk factor for noninvasive respiratory support failure in de novo respiratory failure [23]. Dyspnea is also associated with increased respiratory drive in mechanically ventilated patients, which is itself associated with poor outcomes [24]. Additionally, heterogeneity of treatment response according to dyspnea has been described in COVID-19 patients treated with noninvasive respiratory support [25]. However, whether personalization of noninvasive respiratory support changes patients’ dyspnea, discomfort and outcomes, or whether personalization of noninvasive respiratory support based on patient’s dyspnea or comfort changes outcomes remains to be tested. Unlike a previous study, poor functional status before ICU admission was not associated with ICU mortality [26]. This finding may be explained by differences in the populations analyzed (mostly non-small cell lung cancer vs. a mixed case of immunocompromised patients), by differences in the prevalence of patients with poor performance status before ICU admission (32% vs. 16% in our study) [26], and by the lack of data regarding the impact of functional status on outcomes outside cancer. All in all, in immunocompromised patients admitted to an ICU for acute hypoxemic respiratory failure, factors associated with in-ICU mortality were related to the underlying cause of immunosuppression, to the severity of the acute disease, and to the response to noninvasive respiratory support. Performance status before ICU admission may not be a good predictor of ICU mortality in this setting.
Regarding ICU survivors, we found that solid cancer and poor performance status before ICU admission were associated with poor outcomes at 6 months, independently from the decision to forgo life-sustaining therapies during ICU stay. The decision to forgo life-sustaining therapies is more frequent in immunocompromised than in immunocompetent patients [27]. A few studies have reported that the decision to forgo life-sustaining therapies was associated with poor long-term outcomes in ICU survivors with solid cancer [26, 28]. We confirm that this decision is also an independent predictor of poor long-term outcomes in a mixed-case population of immunocompromised patients. As for ICU mortality, solid cancer was still associated with high mortality after ICU discharge, up to 73% at 12 months according to a recent retrospective single center study [29]. Here, we confirm that solid cancer is a strong predictor of poor outcome at 6 months among ICU survivors. Likewise, poor functional status before ICU admission is a well-established risk factor for long-term mortality in ICU survivors with solid cancer [26, 28], likely because patients with poor performance status are less able to resume cancer treatments [29]. This finding reinforces the importance of interventions before ICU admission to improve cancer patients’ general condition. Interestingly, severity scores at ICU admission and organ failure during ICU stay, such as the need for intubation, were not independently associated with poor outcomes at 6 months among ICU survivors. All in all, our findings suggest that the decision for ICU readmission for immunocompromised patients should be more driven by the functional status before the index ICU admission, than by the intensity of organ failure during the index ICU stay.
Some limitations must be acknowledged. First, the randomized design of the seminal study, with patients selected based on strict inclusion and exclusion criteria, and the unplanned nature of the analysis could have led to selection bias and may have limited the generalizability of our results. However, the baseline characteristics and outcomes of patients are similar to that reported in the largest international observational study [6]. Second, variables included in the logistic regression models were based on data collected in the seminal study. We cannot exclude that data collected at other early timepoints, ECOG performance status at ICU and hospital discharge [29], timing of the decision to forgo life-sustaining therapies, treatment resumption after ICU discharge for cancer patients [26, 30] would have led to different findings. Third, the selection of variables included in the logistic regression models is debatable. For instance, the need for intubation was not included in the analysis of variables associated with ICU mortality because it occurs later than 24 h after ICU admission in about half of patients admitted to the ICU for acute respiratory failure and treated with first-line noninvasive oxygenation strategies [7, 31]. Likewise, SAPS II was not included in the model because its calculation requires 24 h of hindsight after ICU admission [13]. We rather chose early easy-to-assess variables that may be helpful for the clinician when deciding whether or not a patient could benefit from ICU admission. Last, data were collected at inclusion and not at ICU admission. However, the time elapsed between ICU admission and inclusion was very short (less than 15 min in median), suggesting that data collected at inclusion correspond to data at ICU admission.
Conclusion
In immunocompromised patients admitted to the ICU for acute respiratory hypoxemic failure, solid cancer and increased patient discomfort after initiation of noninvasive respiratory support were associated with ICU mortality, independently from the clinical and radiographic severity. After ICU discharge, solid cancer was still associated with poor outcome at 6 months (i.e. death or survival with poor ECOG performance status of 3 or 4) independently of ECOG performance status before ICU admission and the decision to forgo life-sustaining therapies during ICU stay.
Supplementary Information
Abbreviations
- ECOG
Eastern cooperative oncology group
- ICU
Intensive care unit
- FiO2
Fraction of inspired dioxygen
- PaO2
Partial pressure of arterial dioxygen
- PaCO2
Partial pressure of arterial carbon dioxide
- SAPSII
Simplified acute physiology score II
- SOFA
Sequential organ failure assessment
- SpO2
Pulse oximetry
- ARDS
Acute respiratory distress syndrome
Author contributions
MM and RC are the guarantors of the content of the manuscript, including the data and analysis. MM, RC and AWT designed the study. JPF, SE, FP, MD, NT, GP, MaM, DC, AG, JB, ChrG, CV, JD, GL, SJ, AH, JPQ, JD, DB, AWT and RC conducted the study on enrolled patients. MM and RC analyzed and interpreted the data. MM and RC performed statistical analyses and wrote the first draft of the manuscript. RC and AWT revised the first draft of the manuscript. All authors revised and approved the final version of the manuscript. We gratefully thank Jeffrey Arsham for English—editing the manuscript.
Funding information
The seminal study was funded by the ‘Programme Hospitalier de Recherche Clinique Inter-Régional 2015’ of the French Ministry of Health (API15/P/010), and by grants from Le Nouveau Souffle and AADAIRC.
Data availability
The data sets used and/or analyzed during the current study are available from the corresponding author upon reasonable request.
Declarations
Ethics approval and consent to participate
The protocol was approved by the Ethics Committee Ouest III (Poitiers, France) for French centers and by the local ethics committee for the Italian center. Informed consent from patients or their surrogate was obtained orally, with a written record maintained by the investigator.
Consent for publication
All authors have approved the final version of the manuscript.
Prior abstract presentations
Some of the results of this study have been presented as oral communications at the French Intensive Care Society meeting (Congrès Réanimation) in June 2024
Competing interests
RC reports fees and nonfinancial support from Fisher and Paykel Healthcare, Abbvie and Löwenstein Medical, outside the submitted work. JPF reports grants and non-financial support from Fisher & Paykel Healthcare, during the conduct of the study; grants from the French Ministry of Health outside the submitted work; travel expense coverage to attend scientific meetings, personal fees as member of a scientific board and grants from SOS oxygène outside the submitted work; grant from “la bourse du souffle” outside the submitted work. SE reports grants and non-financial support from Fisher & Paykel Healthcare, during the conduct of the study; grants, personal fees and non-financial support from Aerogen Ltd, outside the submitted work. FP reports grants from ALEXION and personal fees from GILEAD, outside the submitted work. NT reports payments of honoraria for lectures, presentations from Fisher and Paykel Healthcare, and support for attending meetings and/or travel from Gilead.Chr G reports travel expense coverage to attend scientific meetings, personal fees and logistic support from Fisher & Paykel, Resmed and Löwenstein Medical, outside the submitted work. SJ reports personal fees for lectures from Hamilton Medical and Nihon Kohden, outside the submitted work.AWT reports grants and personal fees (travel expense coverage to attend scientific meetings and payment for lectures) from Fisher & Paykel, outside the submitted work. MM, MD, GP, MaM, DC, AG, JB, JD, CV, GL, JPQ, JD, DB and SR report no conflict of interest.
Footnotes
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Deceased.
Contributor Information
Mélanie Métais, Email: melanie.metais06@gmail.com.
the FLORALI-IM study group and the REVA Research Network:
Emmanuel Vivier, Saad Nseir, Gwenhaël Colin, Didier Thevenin, Giacomo Grasselli, David Bougon, Mona Assefi, Claude Guérin, Thierry Lherm, and Achille Kouatchet
References
- 1.Rubin LG, Levin MJ, Ljungman P, et al. 2013 IDSA clinical practice guideline for vaccination of the immunocompromised host. Clin Infect Dis. 2013;58(3):e44–100. [DOI] [PubMed] [Google Scholar]
- 2.Quaresma M, Coleman MP, Rachet B. 40-year trends in an index of survival for all cancers combined and survival adjusted for age and sex for each cancer in England and Wales, 1971–2011: a population-based study. Lancet. 2015;385(9974):1206–18. [DOI] [PubMed] [Google Scholar]
- 3.Moreau A-S, Martin-Loeches I, Povoa P, et al. Impact of immunosuppression on incidence, aetiology and outcome of ventilator-associated lower respiratory tract infections. Eur Respir J. 2018. 10.1183/13993003.01656-2017. [DOI] [PubMed] [Google Scholar]
- 4.Vincent J-L, Marshall JC, Namendys-Silva SA, et al. Assessment of the worldwide burden of critical illness: the intensive care over nations (ICON) audit. Lancet Respir Med. 2014;2(5):380–6. [DOI] [PubMed] [Google Scholar]
- 5.Azoulay E, Mokart D, Kouatchet A, Demoule A, Lemiale V. Acute respiratory failure in immunocompromised adults. Lancet Respir Med. 2019;7(2):173–86. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Azoulay E, Pickkers P, Soares M, et al. Acute hypoxemic respiratory failure in immunocompromised patients: the efraim multinational prospective cohort study. Intensive Care Med. 2017;43(12):1808–19. [DOI] [PubMed] [Google Scholar]
- 7.Coudroy R, Frat J-P, Ehrmann S, et al. High-flow nasal oxygen alone or alternating with non-invasive ventilation in critically ill immunocompromised patients with acute respiratory failure: a randomised controlled trial. Lancet Respir Med. 2022;10(7):641–9. [DOI] [PubMed] [Google Scholar]
- 8.Soares M, Toffart A-C, Timsit J-F, et al. Intensive care in patients with lung cancer: a multinational study. Ann Oncol. 2014;25(9):1829–35. [DOI] [PubMed] [Google Scholar]
- 9.Tavares M, Lemiale V, Mokart D, et al. Determinants of 1-year survival in critically ill acute leukemia patients: a GRRR-OH study. Leuk Lymphoma. 2018;59(6):1323–31. [DOI] [PubMed] [Google Scholar]
- 10.West H (Jack), Jin JO, editors. Performance status in patients with cancer. JAMA Oncol. 2015;1(7):998. [DOI] [PubMed]
- 11.Hendriks LE, Kerr KM, Menis J, et al. Non-oncogene-addicted metastatic non-small-cell lung cancer: ESMO clinical practice guideline for diagnosis, treatment and follow-up. Ann Oncol. 2023;34(4):358–76. [DOI] [PubMed] [Google Scholar]
- 12.Gonzalez F, Starka R, Ducros L, et al. Critically ill metastatic cancer patients returning home after unplanned ICU stay: an observational, multicentre retrospective study. Ann Intensive Care. 2023;13:73. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Le Gall JR, Lemeshow S, Saulnier F. A new simplified acute physiology score (SAPS II) based on a European/North American multicenter study. JAMA. 1993;270(24):2957–63. [DOI] [PubMed] [Google Scholar]
- 14.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;22(7):707–10. [DOI] [PubMed] [Google Scholar]
- 15.Bredin S, Decroocq J, Devautour C, Charpentier J, Vigneron C, Pène F. Impact of critical illness on continuation of anticancer treatment and prognosis of patients with aggressive hematological malignancies. Ann Intensive Care. 2024;14(1):143. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Contejean A, Lemiale V, Resche-Rigon M, et al. Increased mortality in hematological malignancy patients with acute respiratory failure from undetermined etiology: a groupe de recherche En Réanimation respiratoire En Onco-Hématologie (Grrr-OH) study. Ann Intensive Care. 2016;6:102. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Azoulay E, Moreau D, Alberti C, et al. Predictors of short-term mortality in critically ill patients with solid malignancies. Intensive Care Med. 2000;26(12):1817–23. [DOI] [PubMed] [Google Scholar]
- 18.Azoulay E, Mokart D, Pène F, et al. Outcomes of critically ill patients with hematologic malignancies: prospective multicenter data from France and Belgium—a groupe de recherche respiratoire en réanimation onco-hématologique study. J Clin Oncol. 2013;31(22):2810–8. [DOI] [PubMed] [Google Scholar]
- 19.Puxty K, McLoone P, Quasim T, Sloan B, Kinsella J, Morrison DS. Risk of critical illness among patients with solid cancers: a population-based observational study. JAMA Oncol. 2015;1(8):1078. [DOI] [PubMed] [Google Scholar]
- 20.Vincent F, Soares M, Mokart D, et al. In-hospital and day-120 survival of critically ill solid cancer patients after discharge of the intensive care units: results of a retrospective multicenter study—a Groupe de recherche respiratoire en réanimation en onco–hématologie (Grrr-OH) study. Ann Intensive Care. 2018;8(1):40. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Lemiale V, Lambert J, Canet E, et al. Identifying cancer subjects with acute respiratory failure at high risk for intubation and mechanical ventilation. Respir Care. 2014;59(10):1517–23. [DOI] [PubMed] [Google Scholar]
- 22.Pham T, Pesenti A, Bellani G, et al. Outcome of acute hypoxaemic respiratory failure: insights from the lung safe study. Eur Respir J. 2021. 10.1183/13993003.03317-2020. [DOI] [PubMed] [Google Scholar]
- 23.Demoule A, Baptiste A, Thille AW, et al. Dyspnea is severe and associated with a higher intubation rate in de novo acute hypoxemic respiratory failure. Crit Care. 2024;28(1):174. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Le Marec J, Hajage D, Decavèle M, et al. High airway occlusion pressure is associated with dyspnea and increased mortality in critically ill mechanically ventilated patients. Am J Respir Crit Care Med. 2024. 10.1164/rccm.202308-1358OC. [DOI] [PubMed] [Google Scholar]
- 25.Grieco DL, Menga LS, Cesarano M, et al. Phenotypes of patients with COVID-19 who have a positive clinical response to helmet noninvasive ventilation. Am J Respir Crit Care Med. 2022;205(3):360–4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Vigneron C, Charpentier J, Wislez M, et al. Short-term and long-term outcomes of patients with lung cancer and life-threatening complications. Chest. 2021;160(4):1560–4. [DOI] [PubMed] [Google Scholar]
- 27.Cortegiani A, Madotto F, Gregoretti C, et al. Immunocompromised patients with acute respiratory distress syndrome: secondary analysis of the LUNG SAFE database. Crit Care. 2018;22(1):157. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Gheerbrant H, Timsit J-F, Terzi N, et al. Factors associated with survival of patients with solid cancer alive after intensive care unit discharge between 2005 and 2013. BMC Cancer. 2021;21(1):9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Benguerfi S, Messéant O, Painvin B, et al. Factors associated with cancer treatment resumption after ICU stay in patients with solid tumors. Ann Intensive Care. 2024;14(1):135. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Borcoman E, Dupont A, Mariotte E, et al. One-year survival in patients with solid tumours discharged alive from the intensive care unit after unplanned admission: a retrospective study. J Crit Care. 2020;57:36–41. [DOI] [PubMed] [Google Scholar]
- 31.Lemiale V, Mokart D, Resche-Rigon M, et al. Effect of noninvasive ventilation vs oxygen therapy on mortality among immunocompromised patients with acute respiratory failure: a randomized clinical trial. JAMA. 2015;314(16):1711–9. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The data sets used and/or analyzed during the current study are available from the corresponding author upon reasonable request.