STRENGTHS AND LIMITATIONS OF THIS STUDY
Prospectively collected data obtained from 44 intensive care units in three countries.
Analytical methods informed by guidelines in pulmonary and critical care, and accounting for competing risks.
Secondary analysis of randomised trial data evaluating the effect of probiotics on incident ventilator-associated pneumonia, without protocolisation of antimicrobial use.
Ventilator-associated pneumonia outcome adjudicated in duplicate in the absence of a universally agreed on definition
Association, but not causation, is examined with these methods.
Introduction
Frailty—conceptualised as a syndrome or as a state of accumulated health deficits—exists on a continuum. Frailty increases an individual’s vulnerability to poor outcomes following a stressor such as critical illness. 1 Numerous studies have documented that critically ill patients with baseline frailty are predisposed to prolonged life support dependency (vasoactive drugs and invasive ventilation), extubation failure, tracheostomy, dialysis, blood transfusions, increased intensive care unit (ICU) and hospital length of stay, and death. 2 , 6 among hospital survivors of critical illness, frailty is associated with impaired health-related quality of life, poor recovery, discharge to a long-term care facility and increased healthcare costs. 7 , 9
The immunosenescence 10 and proinflammatory state 11 12 described in frail individuals is associated with depleted physiological reserve that may increase the risk of developing healthcare-associated infections when critically ill. This dysregulated immunity is one mechanism which may account for worse outcomes among frail hospitalised persons. In a prospective cohort of 203 patients with suspected infection who were ≥75 years old, those who were assessed as frail as defined by a score of 5 or more using the Clinical Frailty Score (CFS) were more likely to develop septic shock (adjusted OR 1.83; 95% CI 1.08 to 2.51) and more likely to die within 30 days of emergency room presentation (OR 2.05; 95% CI 1.02 to 5.24) than those without frailty. 8 In another multicentre cohort of 650 adults with newly suspected infection, including 599 patients with sepsis, those identified as frail were less likely to be discharged home. 13 Among 2784 patients aged ≥65 years, investigators identified frailty using Fried’s phenotype of three or more of the following: weakness, exhaustion, weight loss, low physical activity and low walking speed. 14 In this cohort, which linked regional registries retrospectively over 10 years, frailty was associated with a 78% higher risk of infection-related hospitalisation. 15 In turn, some infections, such as SARS-CoV-2, have in turn been associated with robust upregulation of markers of cellular senescence. 16 The COVID-19 pandemic highlighted the interdependency of frailty, immunity and infections. 17
To examine the relationship between frailty and infectious complications, we will perform a preplanned secondary analysis of a randomised trial database. The PROSPECT trial ( Pro biotics to Prevent S evere P neumonia and E ndotracheal C olonization T rial) enrolled patients in 44 ICUs across Canada, the USA and Saudi Arabia. Eligible patients were expected to be mechanically ventilated for at least 72 hours and were randomised to receive probiotic Lactobacillus rhamnosus GG (1×10 10 colony forming units) or placebo. 18 There was no difference in the primary outcome of ventilator-associated pneumonia (VAP), or any secondary outcomes including ICU-acquired infections , ICU and hospital length of stay or mortality. 19 We also found no effect modification of baseline frailty in a prespecified subgroup analysis, despite our hypothesis that the effect of probiotics may be attenuated in those with a CFS of 5 or more.
The primary objective of the current study is to examine whether baseline frailty is associated with the risk of developing VAP among invasively ventilated adults. We hypothesise that critically ill persons who were frail before hospitalisation would be more likely to develop VAP than persons without frailty.
The secondary objective is to examine whether baseline frailty is associated with the risk of developing other ICU-acquired infections. We hypothesise that critically ill persons who were frail before hospitalisation would be more likely to develop other ICU-acquired infections than persons without frailty.
The tertiary objectives are to examine the relationship between frailty and length of stay, discharge to a long-term care facility and vital status in hospital. We hypothesise that persons with frailty would have a longer length of hospital stay, higher probability of being discharged to a long-term care facility and higher risk of death than persons without frailty.
Methods
This protocol describes a preplanned analysis of the PROSPECT trial database ( NCT02462590 ). We will conduct this preplanned analysis from 1 December 2025 to 31 January 2026.
To prepare for the current study, we conducted a prospective pilot study in which ICU research coordinators completed the CFS on 150 critically ill patients. This showed that research staff can complete CFS scores using medical chart review, with good inter-rater reliability among ICU research coordinators, occupational therapists and geriatrics residents. 20 Frailty assessments were incorporated into the trial as of 23 May 2016, but were not captured retrospectively. Therefore, 2182 out of 2560 enrolled patients (85.2%) will contribute to this substudy.
Research coordinators in all participating centres were trained at each site initiation visit and periodically at study meetings on use of the 9-point CFS by Rockwood and Theou. 21 Frailty assessments were captured by trained research staff and based on patient status prior to the index hospitalisation.
Aside from the experimental intervention of probiotic versus placebo, there was no protocolisation of selective digestive decontamination. We surveyed participating trial sites to determine this at the beginning and end of the trial. By contrast, our site survey showed that all of the sites used chlorhexidine for VAP prevention.
Dependent variables
Primary outcome
The primary outcome will be VAP at any time in the ICU, VAP is defined by the presence of a new, progressive or persistent radiographic infiltrate on chest radiograph after at least 2 days of mechanical ventilation, plus any two of the following: (1) fever (core temperature >38°C) or hypothermia (temperature <36°C); (2) white blood cell count less than 3.0×10 6 /L or exceeding 10×10 6 /L and (3) purulent sputum. Definitions are reported elsewhere. 18 19
Secondary outcomes
Secondary outcomes will be several other ICU-acquired infections, including bloodstream infections (primary or secondary), skin and soft tissue infections, intra-abdominal infections, Clostridioides difficile infections (CDIs) (requiring diarrhoea and laboratory confirmation or colonoscopic or histopathological evidence) and upper genitourinary tract infections, adapted from the International Sepsis Forum. 22 In addition, we will examine the composite of any of the foregoing ICU-acquired infections, as per the main trial report. 19 The isolation of Lactobacillus species in a culture from a sterile site or as the sole or predominant organism in a non-sterile site, defined as an adverse event in the main trial, 19 will be another secondary outcome. We will not include a positive urine culture alone as an infection (as this typically represents asymptomatic bacteriuria), or positive cultures obtained only through microbiological surveillance. Definitions are reported elsewhere. 18 19
Tertiary outcomes
Tertiary outcomes include hospital mortality and the composite of hospital length of stay >60 days or death, and the composite of discharge to a long-term care facility or death.
Adjudication and classification of primary and secondary outcomes
Clinically suspected VAP and CDI were centrally adjudicated using the clinical notes and by microbiological and radiological source reports, following pilot calibration by two independent physicians blinded to allocation and centre. Disagreements were resolved by discussion or by a third investigator. Other incident infections were adjudicated by one physician blinded to allocation and centre.
Clinically suspected infections were classified as prevalent if present before randomisation, on the day of randomisation, or 1 day after randomisation. Prevalent infections were centrally adjudicated by one physician blinded to treatment allocation and centre but were not considered outcomes for the trial. As a result, we will exclude prevalent infections in the current study. Incident infections were those occurring two or more days after randomisation and were considered outcomes for the trial and outcomes for this study.
Independent variables
These baseline variables will be included as independent variables in regression for the analyses of all of the primary, secondary and tertiary outcomes.
Baseline variables
CFS (categorical: CFS >5 versus <5). 23
Age (continuous variable).
Sex (dichotomous variable).
Body mass index (categorical variable: underweight (<18.5 kg/m 2 ), normal weight (18.5–24.9 kg/m 2 ), overweight (25.0–29.9 kg/m 2 ), obese (30.0–39.9 kg/m 2 ) and morbidly obese (≥40.0 kg/m 2 ). 24 25
Acute Physiology and Chronic Health Evaluation II score (continuous variable). 26
Chronic health conditions (one dichotomous variable, any of the following).
Respiratory failure.
Heart failure.
Hepatic failure or cirrhosis.
End-stage renal disease requiring dialysis.
Lymphoma, leukaemia, multiple myeloma or metastatic cancer.
Other immunocompromised state (post chemotherapy, radiotherapy, alcoholism, recent high dose steroids >15 mg/kg for >5 days or steroids over last 30 days).
Probiotics versus placebo (two categories per randomised group).
Analyses of primary and secondary outcomes
We will use Cox proportional hazards regression analysis to assess the association of frailty with our outcomes of interest. We will follow recommendations on control of confounding and reporting of results in causal inference studies of critically ill patients, focusing on association rather than causation. 27 As recommended, variables were not selected based on statistical testing, but rather on the basis of known or suspected association with our outcomes of interest. ICU discharge and death in the ICU will be considered censoring events for all infections measured in the ICU. For CDI, which was collected throughout the hospital stay, censoring events will be hospital discharge, death in the ICU and in the hospital. The proportional hazards assumption will be assessed and handled, if necessary, using time-varying effects or alternative methods if required.
Analyses of tertiary outcomes
To evaluate whether frailty is associated with hospital mortality, we will use Cox proportional hazards regression to assess the relationship between frailty and mortality.
To evaluate whether frailty is associated with prolonged hospital stay >60 days or death, and to evaluate whether frailty is associated with discharge to a long-term care facility or death, we will use logistic regression to assess the relationship between frailty and the composite outcome of prolonged hospital length of stay (>60 days) and/or hospital mortality.
Sensitivity analyses for VAP and any ICU-acquired infection
For the primary outcome of VAP and the secondary outcome of the composite of ICU-acquired infections, we will conduct the following sensitivity analyses:
To further consider the association of infection with level of frailty, we will conduct an analysis including CFS as a continuous linear variable.
To examine the possible effect of a competing risk of death, we will conduct an analysis for the competing risk of death.
Regression results will be presented as adjusted HRs (aHRs) or ORs (aORs) with 95% CIs for the associations of each independent variable with the primary and secondary outcomes, and with the tertiary outcome of hospital mortality. ORs with 95% CIs for the associations of each independent variable with the tertiary composite outcome of prolonged hospital stay or death. Two-tailed p values <0.05 will be considered statistically significant. No corrections for multiple testing will be performed, but findings will be cautiously interpreted in light of CIs.
Patient and public involvement
This preplanned secondary analysis of the PROSPECT Trial was funded by the Canadian Frailty Network (CFN). CFN is a not-for-profit public-facing organisation funded by the Canadian government through the Networks of Centres of Excellence, devoted to improving care for older Canadians living with frailty and supporting their families and caregivers. CFN is dedicated to increasing frailty recognition and assessment, and creating evidence for decision-making, and engaging with older adults and their families and caregivers to disseminate research findings. While we have no specific formal patient or family advisor for this substudy, CFN has several private patient and caregiver partners. CFN cofunded the PROSPECT trial and prioritised additional funding of this preplanned secondary analysis focused on exploring the relationship between frailty and infectious outcomes. We also engaged members of the public before starting the PROSPECT trial in a substudy nested within the PROSPECT Pilot trial in which substitute decision makers were interviewed at the time of patient enrolment to explore their comfort with probiotics. 18 We found that among 103 substitute decision-makers in eight centres, the rationale for their consent related to personal beliefs regarding possible benefits to the patient, as well as predictions of the patient’s wishes regarding this trial; no differences were found in the characteristics of those who consented versus declined the trial. Further, patients in the trial who gained capacity after resolution of their critical illness were asked to agree to ongoing trial participation. When results of this study are available, we will ensure that they are disseminated to members of the public by enlisting the help of CFN patient advisors to translate each presentation or manuscript into suitable tools for written, visual or auditory tools suitable for the public. As citizen awareness about frailty increases, we anticipate strong public interest in this study.
Ethics and dissemination
The original trial was approved by research ethics board and regulatory agencies as relevant in all participating centres in Canada, the USA and Saudi Arabia, prior to any patient enrolment. The Hamilton Integrated Research Ethics Board affiliated with McMaster University was the research board of record for the original trial.
This specific protocol for a secondary analysis of the trial was also approved by the Hamilton Integrated Research Ethics Board on 20 August 2015 (Project ID : 19128).
Free and informed consent was obtained for research participants in this database. For this secondary analysis using existing data with primary consent, the original consent and research ethics approval covers secondary analysis without additional consent.
All data were deidentified at source in participating institutions. Human subjects received no compensation for their participation in this trial.
In the forthcoming report, we will outline baseline characteristics of patients with and without frailty. Baseline demographic and descriptive outcome data (infections, length of stay, vital status) will be reported using descriptive statistics, including means and SDs (or medians with IQRs, if appropriate) for continuous data, and numbers with percentages for categorical data. We will display the day of the first infection over the ICU stay.
This trial was peer-reviewed and funded for the Pilot by the Canadian Institutes of Health Research in January 2013, and for the full trial by the Canadian Institutes of Health Research in July 2016, CFN in March 2017, Physician Services Incorporated in May 2016, Hamilton Academic Health Sciences Organization in January 2016 and Academic Medical Organization of Southwestern Ontario in April 2016. The first patient enrolment of the pilot trial was 22 October 2013 and the final patient was enrolled on 13 March 2019. Overall, 2650 patients were enrolled in this trial. The results of the probiotics trial were published. 19
Findings will be disseminated to investigators at a meeting of the Canadian Critical Care Trials Group. We will present study results at an international conference in the fields of critical care and infectious diseases, to coincide with or precede open-access peer-review publication. To aid knowledge dissemination, we will use a variety of formats. For example, for traditional and social media, we will create two different visual abstracts and infographics of our results suitable to share on clinician-facing and public-facing platforms.
Discussion
In this study, we will examine the association of prehospital frailty, as defined by the CFS, with VAP and other ICU-acquired infections, length of stay and vital status among invasively ventilated critically ill adults, building on existing evidence about the health consequences of frailty. 28 29 This preplanned secondary analysis of the trial database was separately funded by the CFN in September 2024.
Limitations
Limitations of this preplanned secondary study include the risk of confounding and as such, this analysis will yield information on association, not causation. The cointervention of antimicrobials was not protocolised across patients in the PROSPECT trial, and we acknowledge a possible propensity to administer more antimicrobials among patients with and without frailty. Some potential effect modifiers cannot be tested (eg, neutropenia, post-transplantation status, CD4 count <200 cells/uL) because patients with these conditions were excluded from this trial due to concerns about risk of iatrogenic probiotic infection. 18 Diabetes was not recorded, nor were changes to goals of care which could represent a competing risk. We lack translational data about any mechanisms that might help to interpret the findings.
Comparison with prior work
Compared with prior work, this study will include a large contemporary database of invasively ventilated critically ill patients. The Cox model will assess the association of frailty adjusted for other baseline factors (eg, acute illness severity). We will avoid overfitted regression models and conduct sensitivity analyses to evaluate the consistency of results. Half of the patients received probiotics, allowing evaluation of risk factors for infection equally with and without this exposure, which had no effect on the risk of developing VAP in the main trial. 19 Heterogeneous patients from academic and community hospitals will enhance the generalisability of the results.
In summary, this study will identify whether frailty is associated with increased, decreased or no change in risk of ICU-acquired infections in invasively ventilated patients, adjusted for other baseline factors. Results may be useful to patients, their caregivers and clinicians, and could inform practice guidelines and the design of future research.
Acknowledgements
We are grateful to the patients and families participating in PROSPECT (Probiotics: Prevention of Severe Pneumonia and Endotracheal Colonization Trial), as well as the collaborating Research Coordinators and Investigators, and bedside clinicians who supported this work. The trial was designed by the PROSPECT Steering Committee, the PROSPECT Investigators and Research Coordinators and the Canadian Critical Care Trials Group. We are grateful to Dr Stephanie Sibley for her review of this manuscript on behalf of the Canadian Critical Care Trials Group.
Footnotes
Funding: This trial was peer-review funded by the Canadian Frailty Network, the Canadian Institutes of Health Research, the Hamilton Academic Health Sciences Organization and Academic Medical Organization of Southwestern Ontario and the Physician Services Incorporated of Ontario. The manufacturers of L rhamnosus GG (iHealth) donated the blinded study product for the original trial (no funding). FL is a recipient of a Research Career Award from the Fonds de la recherche du Québec-Santé. JCM holds the Unity Health Chair in Trauma Research. DJC holds a Canada Research Chair in Knowledge Translation in Critical Care.
Prepub: Prepublication history for this paper is available online. To view these files, please visit the journal online ( https://doi.org/10.1136/bmjopen-2025-105227 ).
Patient consent for publication: Not applicable.
Provenance and peer review: Not commissioned; externally peer reviewed.
Collaborators: For the PROSPECT Investigators and the Canadian Critical Care Trials Group.
Patient and public involvement: Patients and/or the public were involved in the design, or conduct, or reporting, or dissemination plans of this research. Refer to the Methods section for further details.
References
1. Kim DH, Rockwood K. Frailty in Older Adults. N Engl J Med. 2024;391:538–48. doi: 10.1056/NEJMra2301292.
2. Muscedere J, Waters B, Varambally A, et al. The impact of frailty on intensive care unit outcomes: a systematic review and meta-analysis. Intensive Care Med. 2017;43:1105–22. doi: 10.1007/s00134-017-4867-0.
3. Bagshaw SM, Stelfox HT, McDermid RC, et al. Association between frailty and short- and long-term outcomes among critically ill patients: a multicentre prospective cohort study. CMAJ. 2014;186:E95–102. doi: 10.1503/cmaj.130639.
4. Bagshaw SM, Stelfox HT, Johnson JA, et al. Long-Term Association Between Frailty and Health-Related Quality of Life Among Survivors of Critical Illness. Crit Care Med. 2015;43:973–82. doi: 10.1097/CCM.0000000000000860.
5. Fernando SM, McIsaac DI, Rochwerg B, et al. Frailty and associated outcomes and resource utilization following in-hospital cardiac arrest. Resuscitation. 2020;146:138–44. doi: 10.1016/j.resuscitation.2019.11.011.
6. Fernando SM, McIsaac DI, Rochwerg B, et al. Frailty and invasive mechanical ventilation: association with outcomes, extubation failure, and tracheostomy. Intensive Care Med. 2019;45:1742–52. doi: 10.1007/s00134-019-05795-8.
7. Muscedere J, Bagshaw SM, Kho M, et al. Frailty, Outcomes, Recovery and Care Steps of Critically Ill Patients (FORECAST): a prospective, multi-centre, cohort study. Intensive Care Med. 2024;50:1064–74. doi: 10.1007/s00134-024-07404-9.
8. Fernando SM, Guo KH, Lukasik M, et al. Frailty and associated prognosis among older emergency department patients with suspected infection: A prospective, observational cohort study. CJEM. 2020;22:687–91. doi: 10.1017/cem.2020.377.
9. Cecconi M, Leaver S, Jung C. Caring for frail patients in the ICU: a multidimensional approach. Intensive Care Med. 2024;50:583–6. doi: 10.1007/s00134-024-07364-0.
10. Wilson D, Jackson T, Sapey E, et al. Frailty and sarcopenia: The potential role of an aged immune system. Ageing Res Rev. 2017;36:1–10. doi: 10.1016/j.arr.2017.01.006.
11. Pothier K, Gana W, Bailly N, et al. Associations Between Frailty and Inflammation, Physical, and Psycho-Social Health in Older Adults: A Systematic Review. Front Psychol. 2022;13:805501. doi: 10.3389/fpsyg.2022.805501.
12. Tran Van Hoi E, De Glas NA, Portielje JEA, et al. Biomarkers of the ageing immune system and their association with frailty - A systematic review. Exp Gerontol. 2023;176:112163. doi: 10.1016/j.exger.2023.112163.
13. Komori A, Abe T, Yamakawa K, et al. Characteristics and outcomes of frail patients with suspected infection in intensive care units: a descriptive analysis from a multicenter cohort study. BMC Geriatr. 2020;20:485. doi: 10.1186/s12877-020-01893-1.
14. Fried LP, Tangen CM, Walston J, et al. Frailty in older adults: evidence for a phenotype. J Gerontol A Biol Sci Med Sci. 2001;56:M146–56. doi: 10.1093/gerona/56.3.m146.
15. Trevisan C, Noale M, Amidei CB, et al. Frailty and the risk of infection-related hospitalizations in older age: Differences by sex. Maturitas. 2023;168:1–6. doi: 10.1016/j.maturitas.2022.10.009.
16. Miller WC, Wallace S, Kamm W, et al. Cellular senescence in acute human infectious disease: a systematic review. Front Aging. 2024;5:1500741. doi: 10.3389/fragi.2024.1500741.
17. Vetrano DL, Triolo F, Maggi S, et al. Fostering healthy aging: The interdependency of infections, immunity and frailty. Ageing Res Rev. 2021;69:101351. doi: 10.1016/j.arr.2021.101351.
18. Johnstone J, Heels-Ansdell D, Thabane L, et al. Evaluating probiotics for the prevention of ventilator-associated pneumonia: a randomised placebo-controlled multicentre trial protocol and statistical analysis plan for PROSPECT. BMJ Open. 2019;9:e025228. doi: 10.1136/bmjopen-2018-025228.
19. Johnstone J, Meade M, Lauzier F, et al. for the PROSPECT Investigators and The Canadian Critical Care Trials Group. Effect of Probiotics on Incident Ventilator-Associated Pneumonia in Critically Ill Patients: A Randomized Clinical Trial. JAMA. 2021;326:1–10. doi: 10.1001/jama.2021.13355.
20. Shears M, Takaoka A, Rochwerg B, et al. Assessing frailty in the intensive care unit: A reliability and validity study. J Crit Care. 2018;45:197–203. doi: 10.1016/j.jcrc.2018.02.004.
21. Rockwood K, Theou O. Using the Clinical Frailty Scale in Allocating Scarce Health Care Resources. Can Geriatr J. 2020;23:210–5. doi: 10.5770/cgj.23.463.
22. Calandra T, Cohen J. International Sepsis Forum Definition of Infection in the ICU Consensus Conference. The international sepsis forum consensus conference on definitions of infection in the intensive care unit. Crit Care Med. 2005;33:1538–48. doi: 10.1097/01.CCM.0000168253.91200.83.
23. Ferrante LE, Pisani MA, Murphy TE, et al. The Association of Frailty With Post-ICU Disability, Nursing Home Admission, and Mortality: A Longitudinal Study. Chest. 2018;153:1378–86. doi: 10.1016/j.chest.2018.03.007.
24. Smith RL, Chong TW, Hedrick TL, et al. Does Body Mass Index Affect Infection-Related Outcomes in the Intensive Care Unit? Surg Infect (Larchmt) 2007;8:581–8. doi: 10.1089/sur.2006.079.
25. Serrano PE, Khuder SA, Fath JJ. Obesity as a risk factor for nosocomial infections in trauma patients. J Am Coll Surg. 2010;211:61–7. doi: 10.1016/j.jamcollsurg.2010.03.002.
26. Knaus WA, Draper EA, Wagner DP, et al. APACHE II: a severity of disease classification system. Crit Care Med. 1985;13:818–29.
27. Lederer DJ, Bell SC, Branson RD, et al. Control of Confounding and Reporting of Results in Causal Inference Studies. Guidance for Authors from Editors of Respiratory, Sleep, and Critical Care Journals. Ann Am Thorac Soc. 2019;16:22–8. doi: 10.1513/AnnalsATS.201808-564PS.
28. Rockwood K, Song X, MacKnight C. A global clinical measure of fitness and frailty in elderly people. Can Med Assoc J. 2005;173:489–95. doi: 10.1503/cmaj.050051.
29. Perrella A, Geen O, Ahuja M, et al. Exploring the Impact of Age, Frailty, and Multimorbidity on the Effect of ICU Interventions: A Systematic Review of Randomized Controlled Trials. Crit Care Med. 2024;52:e463–72. doi: 10.1097/CCM.0000000000006315.