Abstract
Background
The aim of this study was to compare the clinical characteristics, treatments and outcomes of immunocompromised patients with viral severe acute respiratory infections (SARI), to those of immunocompetent patients admitted to an intensive care unit (ICU) across Australia.
Methods
Retrospective analysis of a prospective, nation-wide, observational registry of 46 ICUs from June 2022 to August 2025. Critically ill adults (age ≥ 18) with laboratory-confirmed viral SARI were included. Immunocompromised status was defined by the presence of any one of the following: chronic immunosuppression, malignant neoplasm, AIDS/HIV, or organ transplantation. Associations between immunocompromised status and in-hospital mortality were evaluated using mixed-effects logistic regression models. Competing risks regression (Fine and Gray models) was used to assess the association between immunocompromised status and risk of hospital discharge within 30 days.
Results
Among 4,703 patients with viral SARI who were admitted to ICUs, immunocompromised patients accounted for 908 (19.3%) cases. Immunocompromised patients were older (median age 67 vs. 62 years), more comorbid (31.4% vs. 22.5% with ≥ 3 comorbidities), and more frequently admitted with COVID-19 (51.1% vs. 34.5%); (p < 0.001 for all). Immunocompromised patients received more treatment, including antivirals (67.3% vs. 57.7%), and corticosteroids (81.7% vs. 72%) and had higher in-hospital mortality (22.7% vs. 13.1%); (p < 0.01 for all). After adjusting for demographics, comorbidities, vaccination status (where available), ICU interventions and treatments, immunocompromised status was independently associated with nearly double the odds of in-hospital mortality (adjusted OR 1.93, 95% CI 1.57–2.37).
Conclusions
In this study, we found that immunocompromised patients accounted for approximately one-fifth of critically ill viral SARI patients and had nearly twice the odds of in-hospital mortality, when compared to those that were immunocompetent. Targeted public health campaigns and improved treatment strategies may be warranted for this population.
Supplementary Information
The online version contains supplementary material available at https://doi.org/10.1186/s13054-026-06192-6.
Keywords: SARI, ICU, Immunocompromised, Virus
Background
Viral severe acute respiratory infection (SARI) is a major cause of morbidity and mortality worldwide. Even outside pandemic years, SARI generates a vast patient, health care and economic impact, with costs exceeding US$500 billion per year [1]. Understanding how high-risk populations, such as the immunocompromised, are affected by viral SARI is crucial to develop preventative health strategies, improve therapeutic approaches and guide public policy.
This is particularly pertinent as the proportion of critically ill patients with immunosuppression has increased over recent years [2–4]. This group of people includes those on long-term steroids and/or immunosuppressive drugs for autoimmune/inflammatory diseases, those with solid organ and haematological malignancies and people who have undergone solid organ transplants. However, despite the public health importance, there are currently limited high-quality nation-wide data regarding the incidence of viral SARI in immunocompromised hosts, common causative pathogens, ICU management and clinical outcomes.
SPRINT SARI Australia is a national surveillance registry that has collected data from patients with viral SARI admitted to ICUs nation-wide from 2020 to 2025. This platform provides an opportunity to systematically examine the burden of viral SARI in immunocompromised patients across Australia. This study aimed to characterise the incidence, demographics, pathogens, co-morbidities, management and outcomes in immunocompromised patients with viral SARI admitted to ICUs and compare them to immunocompetent individuals.
Methods
Study design and setting
Prospectively collected data from the short-period incidence study of severe acute respiratory infection (SPRINT-SARI) Australia were retrospectively analysed [5, 6]. SPRINT-SARI Australia is a multicentre, observational study of patients admitted with SARI and laboratory-confirmed viral respiratory pathogens, (i.e. including COVID-19 and non-COVID-19) in 46 Australian ICUs. Prior to June 2022, only COVID-19 data were collected, but from this timepoint both COVID-19 and non-COVID-19 SARI cases were recoded. For this study, we only included data collected on or after June 1, 2022 and only from sites after they began reporting both COVID-19 and non-COVID-19 viral SARI cases to ensure consistent case ascertainment and reduce biases related to differences in testing and admission patterns during earlier COVID-19 phases.
Study population
We included all adult patients (aged ≥ 18 years) admitted to ICUs at participating sites with viral SARI due to laboratory-confirmed respiratory pathogens. Individuals were excluded if they were admitted with incidental infection (defined as a positive PCR but did not receive treatment consistent with SARI), or if data concerning immunosuppression status, ICU admission year, and hospital outcomes, were incomplete. Individuals were excluded if they were admitted with incidental infection (defined as a positive PCR but did not receive treatment consistent with SARI because the identified pathogen was not considered the primary reason for ICU admission). We excluded individuals with non-viral SARI including bacterial, fungal, and parasitic co-infections. Characteristics of excluded patients in Supplementary Table 6.
Variables
Immunocompromised status in the SPRINT-SARI Australia dataset followed the APACHE definition by specific diagnoses or treatments that confer significant immune compromise that includes: (1) Hematologic malignancy (leukaemia, lymphoma); (2) Metastatic cancer; (3) HIV/AIDS with severe immunodeficiency; (4) Recent or active chemotherapy for malignancy; (5) Solid-organ transplant on immunosuppressive therapy; (6) Other ongoing immunosuppressive treatment (e.g., high-dose/long-term corticosteroids, cytotoxic drugs, biologic immunosuppressants) (7, 8). Additional variables extracted for analysis included age, sex, ICU admission year, laboratory-confirmed respiratory pathogen primarily responsible for ICU admission, body mass index (BMI), obesity status defined on the basis of a BMI ≥ 30 or by measurement of abdominal girth, number of comorbidities, vaccination status (where available), duration of mechanical ventilation, use of invasive mechanical ventilation, extracorporeal membrane oxygenation (ECMO), renal replacement therapy (RRT), prone positioning, receipt of agents, antibiotics, corticosteroids, and monoclonal antibodies targeting SARS-CoV-2, e.g. Evusheld (tixagevimab/cilgavimab), or IL-6 e.g. tocilizumab during ICU admission, ICU and hospital lengths of stay, ICU discharge location (Supplementary Table 5).
Outcomes
The primary outcome was in-hospital mortality, defined as death in hospital at any time during the index hospital admission. The secondary outcome was hospital discharge within 30 days of ICU admission.
Statistical analysis
Baseline demographics, admission characteristics, interventions and outcomes were compared between the immunocompromised and non-immunocompromised groups using the chi-square or Fisher’s exact tests for categorical variables, and t-tests or Wilcoxon rank-sum tests for continuous variables, as appropriate.
Logistic regression models were used to estimate odds ratios (ORs) and 95% confidence intervals (CIs) for the association between immunocompromised status and in-hospital mortality. To account for clustering of patients within hospitals, mixed-effects logistic regression models were fitted with hospital site as a random effect. Multivariate models were adjusted for covariates based on baseline differences and established clinical predictors of mortality (variables mentioned above). Subgroup analyses were performed to assess the association between immune-competence and in-hospital mortality stratified by type of respiratory pathogen, invasive mechanical ventilation status, and cause of compromised immunity.
To assess the association between immunocompromised status and the likelihood of hospital discharge within 30 days, competing risks regression models (Fine and Gray) were used to estimate sub-distribution hazard ratios (SHRs) and 95% CIs. Time to event was defined as days from ICU admission to hospital discharge (event) or in-hospital death (competing event). Follow-up was administratively truncated at 30 days after ICU admission. Patients who remained hospitalised beyond 30 days were censored at day 30. For these analyses, in-hospital death within 30 days was treated as a competing event. Robust standard errors clustered at the hospital level were used to account for clustering of patients within sites.
For categorical variables with missing data, missing values were coded as a separate category to allow inclusion of all available cases in multivariable analyses. For continuous variables (age), cases with missing data were excluded from multivariable analyses (< 0.07% missing).
Results
Demographics and clinical characteristics
Between 1 June 2022 and 4 August 2025, 4,703 patients met the inclusion criteria from 46 sites (see Fig. 1). The immunocompromised cohort comprised 908 patients (19.3%) while 3,795 patients (80.7%) were non-immunocompromised. The majority of immunocompromised patients had chronic immunosuppression or a malignant neoplasm (Supplementary Table 1). The total number of patients in the database increased each year from 2022 to 2024, with the most recent year comprising 2,169/4,703 (46.1%) of the total cohort.
Fig. 1.
Flow chart of selection of cases
The demographic, admission characteristics, interventions, and outcomes of SARI patients by immunocompromised status are shown in Table 1. The median age was higher in immunocompromised patients (67 years, IQR 57–75) compared with non-immunocompromised patients (62 years, IQR 46–73; p < 0.001). The prevalence of obesity was lower among immunocompromised patients compared with non-immunocompromised patients (11.2% vs. 15.2%, p = 0.01). Immunocompromised patients were more comorbid, with 285 (31.4%) having ≥ 3 comorbidities (e.g. cardiac disease, chronic pulmonary disease, asthma, chronic kidney disease, rheumatologic disorder, liver disease, dementia, chronic neurological disorder, hematologic disease) compared with 855 (22.5%) in the non-immunocompromised group (p < 0.001). Diagnostic categories accounting for an immunocompromised state included chronic immunosuppression (66.3%), malignant neoplasm (52%), organ transplant (12.4%), and HIV/AIDS (2.9%).
Table 1.
Demographic, admission characteristics, interventions and outcomes of SARI patients admitted to ICU according to immunocompromised status
| Immunocompromised status | ||||
|---|---|---|---|---|
| No | Yes | Total | p-value | |
| N | 3,795 (80.7%) | 908 (19.3%) | 4,703 (100.0%) | |
| Age — year | 62 (46–73) | 67 (57–75) | 4,700 | < 0.001 |
| Sex | ||||
| Male | 2,019 (53.2%) | 499 (55.0%) | 2,518 (53.5%) | 0.507 |
| Female | 1,774 (46.7%) | 409 (45.0%) | 2,183 (46.4%) | |
| Not specified | 2 (0.1%) | 0 (0.0%) | 2 (0.0%) | |
| ICU admission year | ||||
| 2022 | 103 (2.7%) | 30 (3.3%) | 133 (2.8%) | 0.767 |
| 2023 | 999 (26.3%) | 237 (26.1%) | 1,236 (26.3%) | |
| 2024 | 1,757 (46.3%) | 412 (45.4%) | 2,169 (46.1%) | |
| 2025 | 936 (24.7%) | 229 (25.2%) | 1,165 (24.8%) | |
| Lab-confirmed respiratory pathogen primarily responsible for ICU admission | ||||
| COVID | 1,309 (34.5%) | 464 (51.1%) | 1,773 (37.7%) | < 0.001 |
| Influenza A virus | 1,186 (31.3%) | 188 (20.7%) | 1,374 (29.2%) | |
| Influenza B virus | 112 (3.0%) | 11 (1.2%) | 123 (2.6%) | |
| Respiratory syncytial virus (RSV) | 322 (8.5%) | 61 (6.7%) | 383 (8.1%) | |
| Parainfluenza virus | 159 (4.2%) | 45 (5.0%) | 204 (4.3%) | |
| Other | 707 (18.6%) | 139 (15.3%) | 846 (18.0%) | |
| Obesity (as defined by clinical staff) | 575 (15.2%) | 102 (11.2%) | 677 (14.4%) | 0.010 |
| Unknown | 29 (0.8%) | 7 (0.8%) | 36 (0.8%) | |
| Number of comorbidities† | ||||
| 0 | 818 (21.6%) | 140 (15.4%) | 958 (20.4%) | < 0.001 |
| 1 | 1,257 (33.1%) | 232 (25.6%) | 1,489 (31.7%) | |
| 2 | 862 (22.7%) | 251 (27.6%) | 1,113 (23.7%) | |
| ≥3 | 855 (22.5%) | 285 (31.4%) | 1,140 (24.2%) | |
| Unknown | 3 (0.1%) | 0 (0.0%) | 3 (0.1%) | |
| COVID-19 vaccination received | ||||
| No | 115 (3.0%) | 26 (2.9%) | 141 (3.0%) | < 0.001 |
| Yes | 840 (22.1%) | 315 (34.7%) | 1,155 (24.6%) | |
| Unknown | 2,840 (74.8%) | 567 (62.4%) | 3,407 (72.4%) | |
| Duration of ventilation (days) | 4 (2–9) | 4 (2–8) | 1,532 | 0.704 |
| Invasive mechanical ventilation | 1,321 (34.8%) | 243 (26.8%) | 1,564 (33.3%) | < 0.001 |
| Unknown | 2 (0.1%) | 0 (0.0%) | 2 (0.0%) | |
| Extracorporeal membrane oxygenation | 84 (2.2%) | 9 (1.0%) | 93 (2.0%) | 0.053 |
| Unknown | 1 (0.0%) | 0 (0.0%) | 1 (0.0%) | |
| Renal replacement therapy (RRT) during ICU admission | 381 (10.0%) | 100 (11.0%) | 481 (10.2%) | 0.428 |
| Unknown | 4 (0.1%) | 0 (0.0%) | 4 (0.1%) | |
| Prone positioning at any time during ICU admission | 155 (4.1%) | 25 (2.8%) | 180 (3.8%) | 0.119 |
| Unknown | 3 (0.1%) | 0 (0.0%) | 3 (0.1%) | |
| Antiviral | 2,189 (57.7%) | 611 (67.3%) | 2,800 (59.5%) | < 0.001 |
| Unknown | 66 (1.7%) | 14 (1.5%) | 80 (1.7%) | |
| Antibiotic | 3,481 (91.7%) | 853 (93.9%) | 4,334 (92.2%) | 0.024 |
| Unknown | 36 (0.9%) | 11 (1.2%) | 47 (1.0%) | |
| Corticosteroid | 2,731 (72.0%) | 742 (81.7%) | 3,473 (73.8%) | < 0.001 |
| Unknown | 40 (1.1%) | 10 (1.1%) | 50 (1.1%) | |
| Monoclonal antibodies (or IL-6 antagonist) § | 71 (1.9%) | 50 (5.5%) | 121 (2.6%) | < 0.001 |
| Unknown | 101 (2.7%) | 34 (3.7%) | 135 (2.9%) | |
| ICU length of stay (days) | 4 (2–7) | 4 (2–7) | 4,698 | 0.425 |
| Hospital length of stay (days) | 10 (6–18) | 12 (6–23) | 4,699 | 0.010 |
| ICU discharge location | ||||
| Death | 363 (9.6%) | 128 (14.1%) | 491 (10.4%) | < 0.001 |
| Wards | 3,094 (81.5%) | 732 (80.6%) | 3,826 (81.4%) | |
| Other hospital | 126 (3.3%) | 28 (3.1%) | 154 (3.3%) | |
| Other rehab | 3 (0.1%) | 2 (0.2%) | 5 (0.1%) | |
| Home | 209 (5.5%) | 18 (2.0%) | 227 (4.8%) | |
| Hospital outcomes | ||||
| In-hospital mortality | 499 (13.1%) | 206 (22.7%) | 705 (15.0%) | < 0.001 |
| 30-day hospital outcomes | ||||
| Hospital discharge within 30 days | 2,977 (78.4%) | 612 (67.4%) | 3,589 (76.3%) | < 0.001 |
| In-hospital death within 30 days | 473 (12.5%) | 187 (20.6%) | 660 (14.0%) | |
| Censored (still hospitalised at day 30) | 345 (9.1%) | 109 (12.0%) | 454 (9.7%) | |
SARI; Severe acute respiratory infection, ICU; Intensive care unit, COVID; Coronavirus disease, IL-6; Interleukin-6, Continuous variables are presented as median (IQR/Interquartile range), † ; Comorbidities include Chronic cardiac disease, including congenital heart disease (not hypertension), Past use of ACE inhibitor or Angiotensin-II receptor blocker, Chronic pulmonary disease (not asthma), Diabetes with complications, Diabetes without complications, Asthma (physician diagnosed), Chronic kidney disease, Rheumatologic disorder, Mild Liver disease, Moderate or severe liver disease, Dementia, Chronic neurological disorder, Chronic hematologic disease, §; Majority received IL-6 antagonists (71%) = 86 patients (85 Tocilizumab + 1 Sarilumab), Antiviral monoclonals = 1 patient (Sotrovimab), Other/unclassified = 34 patients
Over half of the immunocompromised cohort were admitted with COVID-19 (n = 464, 51.1%), followed by influenza A (20.7%) and respiratory syncytial virus (RSV) (n = 61, 6.7%). In contrast, non-immunocompromised patients were more evenly distributed with respect to causative pathogen, with 34.5% admitted with COVID-19, 31.3% with influenza A, and numerically greater proportions of influenza B and RSV compared with immunocompromised patients (Table 1). The top five causes of other types of non-COVID viral respiratory pathogens are listed in the Supplementary Materials 1.
With respect to treatments, antiviral use was more frequent among immunocompromised patients compared with non-immunocompromised patients (67.3% vs. 57.7%, p < 0.001). Similarly, antimicrobials (93.9% vs. 91.7%, p = 0.024), corticosteroids (81.7% vs. 72%, p < 0.001), and monoclonal antibodies (5.5% vs. 1.9%, p < 0.001) were administered more frequently to immunocompromised patients; >70% of administered monoclonal antibodies targeted the IL-6 pathway (i.e. toculizumab or sarilumab).
Overall, one-third of patients required mechanical ventilation (33.3%). Immunocompromised patients were less likely to receive this intervention (26.8% vs. 34.8% p < 0.001) and ECMO was used in 9 immunocompromised patients (1%) compared to 84 (2.2%) non-immunocompromised patients. Prone positioning during ICU admission was reported in 2.8% of patients, with no significant difference between groups. Approximately one in ten patients required RRT, with similar rates in both groups (Table 1).
Outcomes
The median ICU length of stay was 4 days (IQR 2–7), which was the same in both groups. The median length of hospital stay was longer in immunocompromised patients (12 days, IQR 6–23) compared with non-immunocompromised patients (10 days, IQR 6–18) Table 1). ICU mortality was higher in immunocompromised patients (14.1% vs. 9.6%, p < 0.001), and hospital mortality was also greater (22.7% vs. 13.1%, p < 0.001).
We performed univariate and multivariable logistic regression analyses to identify factors associated with in-hospital mortality (Table 2; Fig. 2). After adjusting for covariates, immunocompromised status was associated with nearly double the odds of mortality in the adjusted model (OR 1.93, 95% CI 1.57–2.37). Increasing age was strongly associated with higher mortality, while female patients had lower odds of in-hospital death (OR 0.80, 95% CI 0.67–0.95). Receipt of invasive mechanical ventilation was strongly associated with greater odds of in-hospital mortality (OR 3.67 95% CI 3.02–4.46). Although univariate analysis suggested decreasing mortality over time, this was not confirmed in the multivariable model.
Table 2.
Univariate and multivariate analyses of factors associated with in-hospital mortality
| OR (95% CI) | OR (95% CI)^ | |
|---|---|---|
| Age — year | 1.03 (1.03–1.04) | 1.04 (1.04–1.05) |
| Sex | ||
| Male | Reference | Reference |
| Female | 0.73 (0.62–0.86) | 0.80 (0.67–0.95) |
| ICU admission year | ||
| 2022 | Reference | Reference |
| 2023 | 0.53 (0.34–0.85) | 0.69 (0.41–1.14) |
| 2024 | 0.54 (0.34–0.85) | 0.70 (0.42–1.15) |
| 2025 | 0.58 (0.36–0.92) | 0.78 (0.46–1.31) |
| Lab-confirmed respiratory pathogen primarily responsible for ICU admission | ||
| COVID | 1.43 (1.17–1.75) | 1.66 (1.23–2.26) |
| Influenza A virus | Reference | Reference |
| Influenza B virus | 0.68 (0.37–1.26) | 1.03 (0.51–2.04) |
| Respiratory syncytial virus (RSV) | 1.20 (0.87–1.66) | 0.95 (0.64–1.41) |
| Parainfluenza virus | 1.18 (0.78–1.79) | 0.92 (0.57–1.49) |
| Other | 0.89 (0.68–1.16) | 0.88 (0.63–1.22) |
| Immunocompromised status | ||
| Yes | 1.97 (1.64–2.37) | 1.93 (1.57–2.37) |
| No | Reference | Reference |
| Obesity (as defined by clinical staff) | ||
| No | Reference | Reference |
| Yes | 0.84 (0.66–1.07) | 0.84 (0.64–1.09) |
| Unknown | 1.11 (0.46–2.70) | 0.85 (0.30–2.43) |
| Number of comorbidities† | ||
| 0 | Reference | Reference |
| 1 | 1.28 (0.99–1.65) | 1.18 (0.89–1.57) |
| 2 | 1.66 (1.28–2.15) | 1.21 (0.90–1.63) |
| ≥3 | 2.04 (1.58–2.64) | 1.34 (1.00–1.79) |
| Unknown | 4.57 (0.40–51.92) | 2.95 (0.10–85.46) |
| Invasive mechanical ventilation | ||
| No | Reference | Reference |
| Yes | 3.37 (2.84–3.98) | 3.67 (3.02–4.46) |
| Extracorporeal membrane oxygenation | ||
| No | Reference | Reference |
| Yes | 1.94 (1.19–3.17) | 1.28 (0.73–2.24) |
| Renal replacement therapy (RRT) during ICU admission | ||
| No | Reference | Reference |
| Yes | 4.59 (3.72–5.67) | 3.32 (2.61–4.24) |
| Prone positioning at any time during ICU admission | ||
| No | Reference | Reference |
| Yes | 1.50 (1.03–2.18) | 1.03 (0.68–1.58) |
| Antiviral | ||
| No | Reference | Reference |
| Yes | 1.03 (0.87–1.22) | 0.77 (0.61–0.97) |
| Unknown | 1.61 (0.91–2.83) | 1.50 (0.70–3.23) |
| Antibiotic | ||
| No | Reference | Reference |
| Yes | 2.96 (1.84–4.75) | 1.75 (1.06–2.88) |
| Unknown | 3.83 (1.58–9.27) | 2.29 (0.53–9.89) |
| Corticosteroid | ||
| No | Reference | Reference |
| Yes | 1.92 (1.54–2.37) | 1.33 (1.05–1.68) |
| Unknown | 1.78 (0.80–3.98) | 0.69 (0.16–2.95) |
| Monoclonal antibodies (or IL-6 antagonist) | ||
| No | Reference | Reference |
| Yes | 1.92 (1.25–2.95) | 1.32 (0.82–2.13) |
| Unknown | 1.33 (0.85–2.09) | 1.11 (0.63–1.98) |
OR; Odds ratio, 95% CI; 95% confidence interval, ICU; Intensive care unit, COVID; Coronavirus disease, IL-6; Interleukin-6, ^; Adjusted for covariates (Age, Sex, ICU admission year, Lab-confirmed respiratory pathogen primarily responsible for ICU admission, Immunocompromised status, Obesity, Number of comorbidities, COVID-19 vaccination received, Invasive mechanical ventilation, Extracorporeal membrane oxygenation, Renal replacement therapy (RRT) during ICU admission, Prone positioning at any time during ICU admission, Antiviral, Antibiotic, Corticosteroid, Monoclonal antibodies (or IL-6 antagonist)), COVID-19 vaccination status was included in the multivariable model as a covariate to reduce confounding but is not reported due to high rates of missing data (72.4%), †; Comorbidities include Chronic cardiac disease, including congenital heart disease (not hypertension), Past use of ACE inhibitor or Angiotensin-II receptor blocker, Chronic pulmonary disease (not asthma), Diabetes with complications, Diabetes without complications, Asthma (physician diagnosed), Chronic kidney disease, Rheumatologic disorder, Mild Liver disease, Moderate or severe liver disease, Dementia, Chronic neurological disorder, Chronic hematologic disease
Fig. 2.
Forest plot of adjusted odds ratio of in-hospital mortality
We further examined the association between immunocompromised status (versus non-immunocompromised status) and in-hospital mortality in different sub-groups. Immunocompromise was associated with greater odds of in-hospital mortality in almost all patient categories (Supplementary Table 2). Subtypes of immunocompromised status demonstrated an increased odds of mortality with chronic immunosuppression (OR 1.83, 95% CI 1.44–2.32) and malignant neoplasm (OR 2.22, 95% CI 1.72–2.87, p < 0.001), whereas no significant associations were identified for patients with solid organ transplantation or HIV/AIDS. (Supplementary Table 1). In sensitivity analyses, the association between immunocompromised status and in-hospital mortality remained consistent, including models excluding COVID-19 vaccination status, ICU interventions and treatment variables, and chronic hematologic disease from the comorbidity count (OR range 1.75–1.95; Supplementary Table 4).
Within 30 days of ICU admission, the cumulative incidence of hospital discharge was lower among immunocompromised patients compared with non-immunocompromised patients, with this trend evident within the first week (Fig. 3). Within 30 days of ICU admission, 3,589 patients (76.3%) were discharged from hospital, 660 (14.0%) died in hospital, and 454 (9.7%) remained hospitalised and were therefore censored at day 30. At day 30, 67.4% of immunocompromised patients had been discharged compared with 78.4% of non-immunocompromised patients (Gray’s test p < 0.001).
Fig. 3.
Cumulative incidence of hospital discharge within 30 days with in-hospital death within 30 days as a competing risk, stratified by immunocompromised status
Table 3 shows the un-adjusted and adjusted SHR of hospital discharge within 30 days. In the multivariate analysis, immunocompromise was independently associated with a 29% lower hazard of hospital discharge within 30 days (SHR = 0.71, 95% CI = 0.64–0.80). Older age, increasing number of comorbidities, admission with COVID-19 (versus influenza A), receipt of invasive ventilation, ECMO, RRT, and prone positioning were statistically significantly associated with a reduced likelihood of 30-day hospital discharge.
Table 3.
Unadjusted and adjusted sub-distribution hazard ratios of hospital discharge within 30 days
| SHR [95% CI] | SHR [95% CI]^ | |
|---|---|---|
| Age — year | 0.99 (0.98–0.99) | 0.99 (0.98–0.99) |
| Sex | ||
| Male | Reference | Reference |
| Female | 1.12 (1.02–1.23) | 1.06 (0.97–1.17) |
| Not specified | 3.41 (1.09–10.70) | 0.87 (0.25–3.10) |
| ICU admission year | ||
| 2022 | Reference | Reference |
| 2023 | 1.24 (0.98–1.55) | 1.17 (1.01–1.35) |
| 2024 | 1.29 (0.99–1.68) | 1.20 (1.01–1.42) |
| 2025 | 1.32 (1.03–1.69) | 1.18 (1.00–1.39) |
| Lab-confirmed respiratory pathogen primarily responsible for ICU admission | ||
| COVID | 0.85 (0.77–0.94) | 0.80 (0.73–0.89) |
| Influenza A virus | Reference | Reference |
| Influenza B virus | 0.97 (0.78–1.21) | 0.92 (0.76–1.10) |
| Respiratory syncytial virus (RSV) | 0.98 (0.86–1.11) | 0.96 (0.84–1.09) |
| Parainfluenza virus | 0.91 (0.73–1.13) | 0.98 (0.81–1.19) |
| Other | 1.08 (0.95–1.22) | 0.95 (0.84–1.07) |
| Immunocompromised status | ||
| Yes | 0.73 (0.65–0.83) | 0.71 (0.63–0.80) |
| No | Reference | Reference |
| Obesity (as defined by clinical staff) | ||
| No | Reference | Reference |
| Yes | 0.98 (0.89–1.08) | 1.04 (0.93–1.16) |
| Unknown | 0.89 (0.68–1.16) | 0.96 (0.72–1.30) |
| Number of comorbidities† | ||
| 0 | Reference | Reference |
| 1 | 0.97 (0.87–1.07) | 0.96 (0.88–1.06) |
| 2 | 0.82 (0.73–0.91) | 0.84 (0.77–0.91) |
| ≥3 | 0.75 (0.66–0.86) | 0.82 (0.74–0.91) |
| Unknown | 0.64 (0.16–2.58) | 1.28 (0.35–4.76) |
| Invasive mechanical ventilation | ||
| No | Reference | Reference |
| Yes | 0.41 (0.38–0.44) | 0.42 (0.39–0.45) |
| Unknown | 0.86 (0.46–1.62) | 0.58 (0.48–0.70) |
| Extracorporeal membrane oxygenation | ||
| No | Reference | Reference |
| Yes | 0.22 (0.16–0.31) | 0.35 (0.25–0.51) |
| Unknown | 1.98 (1.77–2.23) | 0.50 (0.29–0.86) |
| Renal replacement therapy (RRT) during ICU admission | ||
| No | Reference | Reference |
| Yes | 0.29 (0.25–0.34) | 0.42 (0.36–0.49) |
| Prone positioning at any time during ICU admission | ||
| No | Reference | Reference |
| Yes | 0.54 (0.44–0.66) | 0.74 (0.62–0.89) |
| Antiviral | ||
| No | Reference | Reference |
| Yes | 0.93 (0.85–1.01) | 1.08 (0.98–1.18) |
| Unknown | 0.77 (0.59–0.99) | 0.83 (0.59–1.15) |
| Antibiotic | ||
| No | Reference | Reference |
| Yes | 0.49 (0.42–0.57) | 0.59 (0.52–0.67) |
| Unknown | 0.47 (0.37–0.61) | 0.70 (0.41–1.20) |
| Corticosteroid | ||
| No | Reference | Reference |
| Yes | 0.76 (0.69–0.83) | 0.99 (0.90–1.08) |
| Unknown | 0.72 (0.57–0.91) | 0.75 (0.43–1.30) |
| Monoclonal antibodies (or IL-6 antagonist) | ||
| No | Reference | Reference |
| Yes | 0.57 (0.47–0.70) | 0.70 (0.57–0.85) |
| Unknown | 0.89 (0.75–1.07) | 1.07 (0.86–1.33) |
SHR; Sub-distribution hazard ratio, 95% CI; 95% confidence interval, ICU; Intensive care unit, COVID; Coronavirus disease, IL-6; Interleukin-6, ^; Adjusted for covariates (Age, Sex, ICU admission year, Lab-confirmed respiratory pathogen primarily responsible for ICU admission, Immunocompromised status, Obesity, Number of comorbidities, COVID-19 vaccination received, Invasive mechanical ventilation, Extracorporeal membrane oxygenation, Renal replacement therapy (RRT) during ICU admission, Prone positioning at any time during ICU admission, Antiviral, Antibiotic, Corticosteroid, Monoclonal antibodies (or IL-6 antagonist)), COVID-19 vaccination status was included in the multivariable model as a covariate to reduce confouding but is not reported due to high rates of missing data (72.4%), †; Comorbidities include Chronic cardiac disease, including congenital heart disease (not hypertension), Past use of ACE inhibitor or Angiotensin-II receptor blocker, Chronic pulmonary disease (not asthma), Diabetes with complications, Diabetes without complications, Asthma (physician diagnosed), Chronic kidney disease, Rheumatologic disorder, Mild Liver disease, Moderate or severe liver disease, Dementia, Chronic neurological disorder, Chronic hematologic disease
Discussion
Immunocompromised patients accounted for nearly 20% of all ICU admissions to SPRINT-SARI sites in Australia, suggesting this group is over-represented given estimates of immunocompromise in the general population is 2–3% [9]. Immunocompromised patients were older and had higher levels of comorbidities, reflecting increased medical complexity. Immunocompromised patients experienced nearly double the odds of in-hospital mortality compared to non-immunocompromised patients even after adjusting for demographics, comorbidities, vaccination status, ICU interventions and treatments. Immunocompromised patients received antiviral therapies more frequently, which likely reflects differences in pathogen distribution and treatment indications, given the higher prevalence of COVID-19 and influenza A infections in this group.
These findings are consistent with previous studies that demonstrated increased mortality among immunocompromised patients with respiratory infections including virus-detected acute respiratory failure, ventilator-associated lower respiratory tract infection and severe COVID-19 in the ICU [10–15]. In our study, immunocompromised patients accounted for 19% of ICU admissions for viral SARI, consistent with previous studies reporting 12% to 30% for severe respiratory infections [10–15].
There are several potential reasons why immunocompromised patients may have higher mortality. Firstly, they may be more vulnerable to different infections – opportunistic, health-care associated and multi-resistant pathogens (viruses, bacteria and fungi) that don’t respond as well to treatment [16]. Second, the immunocompromised status means innate, or adaptive immune responses are likely to be impaired resulting in higher pathogen burden and slower clearance. Thirdly, immunocompromised patients were older and more comorbid, with other diseases that may impact their host response and ability to combat infection, as well as underlying chronic diseases or transplantation, which reduce physiological reserve and increase the risk of ICU-related complications. Interestingly, the rate of invasive mechanical ventilation was lower in the immunocompromised population (despite similar markers of severity at presentation), which may reflect systematic differences in management, including withholding a treatment because of perceived worse prognosis or higher complications, or due to treatment limitations.
Sub-group analyses indicated that the influence of immunocompromise on mortality was consistent across major respiratory pathogens, including COVID-19 and non-COVID-19 (influenza A, RSV, and other types), suggesting that underlying immune dysfunction rather than the specific pathogen was the key driver of poor outcomes. This aligns with a previous study, which reported that mortality did not vary much across types of respiratory virus. We also found that the association between immunocompromise and mortality persisted regardless of invasive mechanical ventilation, consistent with the prior study [17, 18]. Thus, these results reinforce the consistent, adverse influence of impaired immunity in SARI.
Immunocompromised patients had longer hospital stay compared to non-immunocompromised counterparts suggesting higher healthcare resources usage, as previously reported [19–21]. However, the duration of mechanical ventilation and ICU length of stay were similar. Steroids were commonly administered in both groups. As the majority of patients had COVID-19, it is probable such treatment was guided by the respiratory pathogen isolated, rather than immune status alone. Similarly, the greater utilization of IL-6 antagonists in the immunocompromised group, is likely related to the greater proportion of COVID-19 infections.
The strengths of this study are that it is one of the first nation-wide studies to investigate the impact that immunocompromised status has on outcomes of patients admitted to ICU with viral SARI. The study captures data from over 4700 patients, across 46 sites in Australia, ensuring broad geographic representation and generalizability to similar high-income healthcare settings. It spans multiple respiratory virus seasons, providing robust and recent population-level estimates of incidence and outcomes. Our study includes detailed clinical data including causative pathogens, treatments, interventions, and outcomes, enabling comprehensive adjusted analyses to identify independent risk factors for mortality.
However, we acknowledge several limitations. While a broad range of immunocompromised conditions were captured, the SPRINT SARI Australia database did not record detailed information on the degree or duration of immunosuppression, specific immunosuppressive regimens, or recent changes in therapy, all of which may influence outcomes and therefore lacked granularity. We also acknowledge that immunocompromised status encompasses a heterogeneous spectrum of conditions associated with varying mechanisms and degrees of immune dysfunction. The study also lacked detailed information on comorbidity severity, and whether ACE inhibitor / AR2 blocker use reflected current or historical use. Additionally, microbiological testing practices may have varied across sites and over time, potentially affecting pathogen detection rates and the proportion of cases without identified causative organisms. In addition, we included patients with viral infections as the primary cause and excluded patients with secondary and/or opportunistic non-viral infections, which we acknowledge introduces selection bias. Our work was focussed on the relationship between immunocompromised status and the primary viral infection. While we adjusted for multiple confounders, residual confounding from unmeasured variables such as frailty, functional status, or goals of care discussions cannot be excluded. Missing categorical data were modelled as separate unknown categories to minimise data loss; however, this approach may introduce bias if missingness was not completely random. There was no illness severity score, such as Acute Physiology and Chronic Health Evaluation (APACHE) or Sequential Organ Failure Assessment (SOFA), as well as no information on non-invasive mechanical ventilation or degree of hypoxia in this study. In addition, details regarding the specific type of chronic immunosuppression or malignant neoplasm were not part of the database collection and there is significant overlap between the immunocompromised categories. We also had a small sample size for HIV-infection with no further information on HIV/AIDS status including CD4 + T cell count or treatment. Notably, transplant subgroups and vaccination data were missing in > 70% of the patient population. Subgroups such as HIV/AIDS, transplant recipients, influenza B and parainfluenza groups were noticeably smaller, and should be considered exploratory and hypothesis-generating rather than definitive estimates of effect.
This study may be affected by selection bias arising from differences in referral and ICU admission practices. For the same level of symptoms and clinical observations prior to ICU admission, immunocompromised patients are more likely to be referred for hospital care and more likely to be admitted to ICU than immunocompetent patients. ICU admission is generally expected to improve outcomes for admitted patients, and this benefit is unlikely to be lower in immunocompromised individuals. As a result, immunocompromised patients in this ICU cohort may represent a relatively less severe subset of all immunocompromised patients with severe acute respiratory infection. Notably, immunocompromised patients may be admitted at different clinical thresholds than immunocompetent patients, which complicates interpretation of clinical outcomes and mortality. Finally, SPRINT-SARI Australia captures only patients admitted to ICU, thereby vulnerable to selection bias related to ICU admission practices, potentially missing immunocompromised patients who died before ICU admission. ICU admission itself is not a biologically neutral selection process.
Future research should focus on identifying specific immunocompromised subgroups at highest mortality, risk, evaluating novel therapeutic approaches in randomized controlled trials conducted in these populations, and developing risk stratification tools to guide clinical decision-making and resource allocation. Ultimately, improving outcomes for immunocompromised patients with SARI will require a multifaceted approach encompassing enhanced prevention, early recognition, tailored management strategies, and healthcare system preparedness to meet the complex needs of this growing patient population.
Conclusion
In conclusion, immunocompromised patients with viral SARI were relatively common and represent a high-risk population, with significantly elevated mortality rates and resource utilisation compared to non-immunocompromised patients, even in highly resourced settings. Despite advances in supportive care and the availability of targeted therapies, outcomes remain poor in this vulnerable population. These findings underscore the need for further prevention strategies and the development of evidence-based treatment guidelines specifically designed for immunocompromised patients with SARI.
Supplementary Information
Below is the link to the electronic supplementary material.
Acknowledgements
We would like to thank the SPRINT-SARI Australia Management Committee and participating Australian SPRINT-SARI sites for their support and contributions: Aidan Burrell, Andrew A Udy, Allen Cheng, Tony Trapani, Sze Jye Ng, Shailesh Bihari, Cindy Liang, Kerry Johnson, Jenna Hassall, Winston Cheung, Husna Begum, Mark Plummer, Mark Kol, Ed Litton, Richard Mcallister, Jenipher Alexandria, Mahesh Ramanan, Lewis Campbell, Elissa Milford, Christopher R Andersen, Morgan Rose, Peinan Zhao, Anis Chaba, Aaliya Ibrahim, Simon Erickson, Sherene Magana Cruz, Simon Cumming, Win Wah, Steve McGloughlin.
Abbreviations
- ANZIC-RC
Australia and New Zealand Intensive Care Research Centre
- AIDS
Acquired Immunodeficiency Syndrome
- APACHE
Acute Physiology and Chronic Health Evaluation
- BMI
Body Mass Index
- ECMO
Extracorporeal Membrane Oxygenation
- HIV
Human Immunodeficiency Virus
- ICU
Intensive Care Unit
- RRT
Renal Replacement Therapy
- SARI
Severe Acute Respiratory Infections
- SOFA
Sequential Organ Failure Assessment
- SPRINT SARI
Short Period Incidence Study of Severe Acute Respiratory Infections
Author contributions
A.B, S.C.S, C.R.A and P.H conceptualised and developed the research aims and hypothesis. W. W and A.B conducted the statistical analysis and generated the figures and tables. P.H collated and managed the documents. A.A.U, W.W, A.B, S.C.S, C.R.A, S.J.N provided detailed feedback that was incorporated into the manuscript.
Funding
We would like to acknowledge the funding from Department of Health, Common Wealth of Australia.
Data availability
No datasets were generated or analysed during the current study.
Declarations
Ethics approval and consent to participate
Data were collected using the standardised SPRINT-SARI case report form and managed in a secure, web-based RedCap data management system [22, 23]. Approval for data collection with a waiver of informed consent was provided by the National Mutual Acceptance scheme and Human Research Ethics Committee of Alfred Health (HREC/16/Alfred/59) [6, 24].
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Priyanka Hastak and Christopher R. Andersen contributed equally to this work.
Sarah C. Sasson and Aidan Burrell supervised the work equally.
References
- 1.World Bank Group. From Panic and Neglect to Investing in Health Security: Financing Pandemic Preparedness at a National Level (English) 2025 [Available from: https://documents.worldbank.org/en/publication/documents-reports/documentdetail/979591495652724770/from-panic-and-neglect-to-investing-in-health-security-financing-pandemic-preparedness-at-a-national-level
- 2.Vigneron C, Charpentier J, Valade S, Alexandre J, Chelabi S, Palmieri L-J, et al. Patterns of ICU admissions and outcomes in patients with solid malignancies over the revolution of cancer treatment. Ann Intensiv Care. 2021;11(1):182. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Kerekes DM, Frey AE, Prsic EH, Tran TT, Clune JE, Sznol M, et al. Immunotherapy Initiation at the End of Life in Patients With Metastatic Cancer in the US. JAMA Oncol. 2024;10(3):342–51. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Shah M, Khalid A, Standring O, Amini N, Demyan L, Koti S, et al. Disparities in Immunotherapy Use Prior to First-Line Approval for Metastatic Hepatocellular Carcinoma in the United States. Onco. 2025;5(1):6. [Google Scholar]
- 5.Tang Y, Zhao P, Cheng AC, Ibrahim A, Hassall J, Litton E, et al. Respiratory syncytial virus infection in adult and paediatric patients admitted to intensive care in Australia: A nation-wide comparison with COVID-19. Crit Care Resusc. 2025;27(2):100113. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Otto M, Britton PN, Serpa Neto A, Erickson S, Festa M, Crawford NW, et al. COVID-19 related ICU admissions in paediatric and young adult patients in Australia: a national case series 2020–2022. The Lancet Regional Health – Western Pacific. 2023;36. [DOI] [PMC free article] [PubMed]
- 7.Cheng GS, Crothers K, Aliberti S, Bergeron A, Boeckh M, Chien JW, et al. Immunocompromised Host Pneumonia: Definitions and Diagnostic Criteria: An Official American Thoracic Society Workshop Report. Ann Am Thorac Soc. 2023;20(3):341–53. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Australian Government Department of Health and Aged Care Australian Technical Advisory Group on Immunisation (ATAGI). Australian Immunisation Handbook 2022 [Available from: immunisationhandbook.health.gov.au.
- 9.Harpaz R, Dahl RM, Dooling KL. Prevalence of Immunosuppression Among US Adults, 2013. JAMA. 2016;316(23):2547–8. [DOI] [PubMed] [Google Scholar]
- 10.Vaid N, Ardissino M, Reed TAN, Goodall J, Utting P, Miscampbell M, et al. Clinical characteristics and outcomes of immunosuppressed patients hospitalized with COVID-19: experience from London. J Intern Med. 2021;289(3):385–94. [DOI] [PubMed] [Google Scholar]
- 11.Andersen KM, Bates BA, Rashidi ES, Olex AL, Mannon RB, Patel RC, et al. Long-term use of immunosuppressive medicines and in-hospital COVID-19 outcomes: a retrospective cohort study using data from the National COVID Cohort Collaborative. Lancet Rheumatol. 2022;4(1):e33–41. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Li C, He H, Wang Y, Huang L, Chen Z, Zhang Q, et al. Outcomes and inflammation changes in different types of immunocompromised patients with critically ill COVID-19 admitted to ICU: a national multicenter study. BMC Pulm Med. 2024;24(1):548. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Stainer A, Amati F, Sotgiu G, Puci MV, Nigro M, Bussini L, et al. In-Hospital mortality for immunocompromised patients with SARS-CoV-2 community-acquired pneumonia (CAP). Eur Respir J. 2024;64(68):5081. [Google Scholar]
- 14.Dumas G, Bertrand M, Lemiale V, Canet E, Barbier F, Kouatchet A, et al. Prognosis of critically ill immunocompromised patients with virus-detected acute respiratory failure. Ann Intensive Care. 2023;13(1):101. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Moreau AS, Martin-Loeches I, Povoa P, Salluh J, Rodriguez A, Thille AW et al. Impact of immunosuppression on incidence, aetiology and outcome of ventilator-associated lower respiratory tract infections. Eur Respir J. 2018;51(3):1701656. [DOI] [PubMed]
- 16.Liu C, Rosen EA, Stohs EJ, Imlay H, Nigo M, Gottesdiener LS et al. Tackling antimicrobial resistance in people who are immunocompromised: leveraging diagnostic and antimicrobial stewardship. The Lancet Infectious Diseases. 2026;26(1):e30-e48. [DOI] [PMC free article] [PubMed]
- 17.Lemiale V, Resche-Rigon M, Azoulay E. Early non-invasive ventilation for acute respiratory failure in immunocompromised patients (IVNIctus): study protocol for a multicenter randomized controlled trial. Trials. 2014;15:372. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.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]
- 19.Yang J, Andersen KM, Rai KK, Tritton T, Mugwagwa T, Reimbaeva M, et al. Healthcare resource utilisation and costs of hospitalisation and primary care among adults with COVID-19 in England: a population-based cohort study. BMJ Open. 2023;13(12):e075495. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Burrell AJ, Broadley T, Udy AA. Outcomes for patients with COVID-19 admitted to Australian intensive care units during the first four months of the pandemic. Med J Aust. 2021;215(10):485. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Moradi S, Najafpour Z, Cheraghian B, Keliddar I, Mombeyni R. The Extra Length of Stay, Costs, and Mortality Associated With Healthcare-Associated Infections: A Case-Control Study. Health Sci Rep. 2024;7(11):e70168. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Harris PA, Taylor R, Thielke R, Payne J, Gonzalez N, Conde JG. Research electronic data capture (REDCap)--a metadata-driven methodology and workflow process for providing translational research informatics support. J Biomed Inf. 2009;42(2):377–81. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Harris PA, Taylor R, Minor BL, Elliott V, Fernandez M, O’Neal L, et al. The REDCap consortium: Building an international community of software platform partners. J Biomed Inf. 2019;95:103208. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Burrell AJ, Pellegrini B, Salimi F, Begum H, Broadley T, Campbell LT, et al. Outcomes for patients with COVID-19 admitted to Australian intensive care units during the first four months of the pandemic. Med J Aust. 2021;214(1):23–30. [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
No datasets were generated or analysed during the current study.