Jung et al. Crit Care (2021) 25:149 https://doi.org/10.1186/s13054-021-03551-3 RESEARCH Open Access The impact of frailty on survival in elderly intensive care patients with COVID‑19: the COVIP study Christian Jung1* , Hans Flaatten2,3, Jesper Fjølner4, Raphael Romano Bruno1, Bernhard Wernly5, Antonio Artigas6, Bernardo Bollen Pinto7, Joerg C. Schefold8, Georg Wolff1, Malte Kelm1, Michael Beil9, Sigal Sviri9, Peter Vernon van Heerden10, Wojciech Szczeklik11, Miroslaw Czuczwar12, Muhammed Elhadi13, Michael Joannidis14, Sandra Oeyen15, Tilemachos Zafeiridis16, Brian Marsh17, Finn H. Andersen18,19, Rui Moreno20, Maurizio Cecconi21, Susannah Leaver22, Ariane Boumendil23,24, Dylan W. De Lange25 and Bertrand Guidet23,24 on behalf of COVIP study group Abstract Background: The COVID-19 pandemic has led highly developed healthcare systems to the brink of collapse due to the large numbers of patients being admitted into hospitals. One of the potential prognostic indicators in patients with COVID-19 is frailty. The degree of frailty could be used to assist both the triage into intensive care, and decisions regarding treatment limitations. Our study sought to determine the interaction of frailty and age in elderly COVID-19 ICU patients. Methods: A prospective multicentre study of COVID-19 patients ≥ 70 years admitted to intensive care in 138 ICUs from 28 countries was conducted. The primary endpoint was 30-day mortality. Frailty was assessed using the clinical frailty scale. Additionally, comorbidities, management strategies and treatment limitations were recorded. Results: The study included 1346 patients (28% female) with a median age of 75 years (IQR 72–78, range 70–96), 16.3% were older than 80 years, and 21% of the patients were frail. The overall survival at 30 days was 59% (95% CI 56–62), with 66% (63–69) in fit, 53% (47–61) in vulnerable and 41% (35–47) in frail patients (p < 0.001). In frail patients, there was no difference in 30-day survival between different age categories. Frailty was linked to an increased use of treatment limitations and less use of mechanical ventilation. In a model controlling for age, disease severity, sex, treat- ment limitations and comorbidities, frailty was independently associated with lower survival. Conclusion: Frailty provides relevant prognostic information in elderly COVID-19 patients in addition to age and comorbidities. Trial registration Clinicaltrials.gov: NCT04​321265, registered 19 March 2020. Keywords: COVID-19, Frailty, Outcome, Elderly, Pandemia Introduction The first wave of the SARS-CoV-2 coronavirus dis- *Correspondence: christian.jung@med.uni-duesseldorf.de 1 ease (COVID-19) pandemic disproportionally affected Department of Cardiology, Pulmonology and Vascular Medicine, Medical Faculty, Heinrich-Heine-University Duesseldorf, Moorenstraße 5, the elderly population, creating an unprecedented 40225 Duesseldorf, Germany influx of patients into hospital and intensive care [1]. Full list of author information is available at the end of the article © The Author(s) 2021. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://​creat​iveco​mmons.​org/​licen​ses/​by/4.​0/. The Creative Commons Public Domain Dedication waiver (http://​creat​iveco​ mmons.​org/​publi​cdoma​in/​zero/1.​0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Jung et al. Crit Care (2021) 25:149 Page 2 of 9 Consequently, ICU capacity had to be increased several see Additional file 1, for a map of participating ICUs fold [2]. Despite this, many countries struggled with lim- and patients included see Additional file 2). The study ited resources and were forced into a stricter admission was planned and conducted by the very old intensive policy to ICUs. For various reasons, this disproportion- care patients (VIP) project within the European Society ately affected the very old subgroup of patients, with of Intensive Care Medicine (ESICM) (www.​vipst​udy.​org) reports of ad hoc criteria used for ICU admission, some who also endorsed the study. National coordinators were even based on age alone. responsible for the recruitment of ICUs, coordinated Over the last few years, the assessment of frailty at national and local ethical permission and supervised admission to the ICU has become increasingly popu- patient recruitment at the national level. Ethical approval lar. The clinical frailty scale (CFS) proved to be a useful was mandatory for study participation in each coun- tool in predicting the chances of ICU survival in very old try. Due to the diversity of ethical consent procedures, intensive care patients [3–5] and is one of a number of some countries could recruit patients without informed tools commonly used to assess frailty. In previous studies, consent while the rest had to obtain it. The study delib- the CFS demonstrated a high inter-rater reliability [6]. erately allowed for co-enrolment of study patients to Indeed, the UK National Institute for Health and additional observational COVID-19 studies. To limit Care Excellence (NICE) advocates the use of the CFS workload, screening failures for the study were not in clinical decision making for patients with COVID-19 recorded. The study was registered on ClinicalTrials.gov ≥ 65 years [7]. Additionally, during the present COVID- (ID: NCT04321265) and adhered to the European Union 19 pandemic frailty has been found to be strongly asso- General Data Privacy Regulation (GDPR) directive, ciated with all-cause mortality risk in hospitalised older which is implemented in most participating countries. patients [8, 9]. Hewitt et al. showed in a study of 1564 It was agreed that the first period of the COVIP study adult (> 18 years) COVID-19 patients that frailty was would stop on 26th of May 2020 corresponding with the a better tool for risk stratification than age or concomi- slowing down of the first wave of critically ill patients in tant diseases [10]. In another, smaller study of 677 older most of the participating countries. However, the study (> 65 years) patients with COVID-19, frailty was associ- continued to recruit in order to catch a possible second ated with mortality after a mean follow-up of 34 days [8]. wave. The present study included patients from the first A similar association between frailty and hospital mor- recruitment period. Each participating ICU included tality was shown in a smaller cohort of 42 COVID-19 consecutive patients up to and including those admitted patients [9]. Bellelli et al. also demonstrated in a cohort of on 26th of May 2020. COVID-19 diagnosis was based on 105 patients that frailty could be used for risk evaluation a positive polymerase chain reaction (PCR) test. Patients of COVID-19 and proposed the systematic collection of were followed up until death, 30 days, and three months frailty in all patients at hospital admission [11]. By con- after ICU admission. Similar to the previous VIP studies, trast, recently published retrospective studies focusing a website was set up to facilitate dissemination of infor- on non-ICU COVID-19 patients found that frailty had mation about the study and to allow for data entry using no or low diagnostic or prognostic value [12, 13]. There an electronic case report form (CRF). is therefore need for more high-quality data on patient outcomes to determine whether frailty, as assessed by Study population the CFS, and in particular its interaction with age can be Patients who were 70 years or older with proven COVID- used for prognostication in critically ill elderly patients 19 and admitted to an ICU were eligible. Pre-ICU triage with COVID-19 [14]. This is of paramount importance was not a part of this study. To avoid duplication caused in order to establish an ethical and medically appropriate by the transfer of a patient from one ICU to another, rationing of ICU care. each patient could only be entered once into the database The main aim of the present study was to study the out- regardless of readmission, transfer or other reason. This come of elderly patients with COVID-19 admitted to an resulted in a single electronic CRF per patient. The refer- intensive care unit and to study the influence of frailty on ence date was day 1 of the first admission to an ICU. All outcome. consecutive dates were numbered sequentially from the admission date. Methods Design and setting Data collection This is a prospective multicentre study of COVID-19 Centres collected the data using a uniform online CRF. patients ≥ 70 years old admitted to the ICU. Recruitment The sequential organ-failure assessment (SOFA) score took place from 19 March to 26 May 2020, in 138 inten- on admission was calculated either manually or using sive care units in 28 countries (for a list of collaborators, an online calculator in the electronic CRF as described Jung et al. Crit Care (2021) 25:149 Page 3 of 9 previously [3, 4]. Additionally, the first arterial blood gas unknown. Survival between groups was compared using analysis with ­pO2 [mmHg] and the ­FiO2 [%] to calculate the log-rank test. Incidence of organ support and treat- the ­PaO2/FiO2 index ­(pO2/FiO2 ratio) on admission was ment limitation were estimated using cumulative inci- recorded. Length of stay (LOS) was recorded in hours. As dence analysis considering ICU death and ICU discharge described previously, [3] the electronic CRF and database as competing risks. Comparisons between groups were ran on a secure server set up by and stored at Aarhus performed using Gray’s test. University, Aarhus, Denmark. Multivariate analysis of primary and secondary out- comes: to account for the multilevel structure of the data Frailty and comorbidities with individuals nested into the ICU, all multivariate The frailty level prior to the acute illness and hospital models were built including a random intercept by ICU, admission was assessed using the clinical frailty scale assuming a Gaussian distribution for the random effect. (CFS). This is an intuitive pictographic description along The random effect was tested by comparing log-likeli- with information required to perform the assessment hood of two models including frailty with and without [15]. The CFS defines nine classes from very fit to ter- random effect. minally ill (Additional file 3). The required information Three sequential random effects, multilevel Cox regres- could be obtained either from the patient, the caregiver/ sion models, were used to evaluate the impact of frailty family or hospital records [4, 16]. We used the English on both 30-day and 90-day survival. First, we estimated version of the CFS. Patients with a CFS of 1–3 were clas- the impact of frailty on outcome without adjustment for sified as fit, those with a CFS of 4 as vulnerable and a CFS confounding using a baseline model including only frailty of 5 or higher as frail. CFS assessment was performed as (model 1). Second, to estimate the impact of frailty when described previously with excellent inter-rater variation adjusting for patients’ baseline characteristics, the fol- [4]. The definitions of pre-existing comorbidities are pro- lowing covariates were added to model 1: age (as a con- vided in Additional file 4. tinuous variable), sex, comorbidities, SOFA score, BMI, ­PaO2/FiO2 (as continuous variables). Third, to evalu- Outcome measurement ate whether the effect of frailty was independent of ICU The primary endpoint was the survival status assessed management strategies, both organ support and treat- at 30 days after ICU admission. The outcome at 90 days ment limitation (model 3) were added to model 2 as time- was also assessed. Data could be retrieved either directly, dependent covariates (variables start at 0 for all subjects from the hospital administration system or following tel- and are recoded to 1 only when organ support is received ephone follow-up. Limitation of life-sustaining therapies or when limitation occurs). Two sequential random such as withholding or withdrawing organ support was effects, multilevel cause-specific Cox regression mod- documented based on international recommendations els, were used to evaluate the impact of frailty on organ [17] although there is a large variation in Europe in the support and treatment limitation. First, we estimated the use of end-of-life care [16]. The definitions of organ sup- impact of frailty on variable of interest without adjust- port are detailed in Additional file 5. ment for confounding using a baseline model including only frailty (model 1). Second, to estimate the impact of Statistical analysis frailty when adjusting for patients’ baseline character- No formal sample size calculation prior to this purely istics the following covariates were added to model 1: observational study was performed. The analysis plan age (as a continuous variable), sex, comorbidities, SOFA was finalised prior to any analysis. The primary exposure score, BMI, ­PaO2/FiO2 (as continuous variables). was frailty (fit, vulnerable or frail at ICU admission), the A sensitivity analysis was conducted to investigate primary outcome was 30-day survival, and the secondary whether results differ including only European patients. outcomes were overall survival up to 90 days after ICU All p values were two-sided, and p < 0.05 was consid- admission, organ support (vasoactive drugs, mechanical ered statistically significant. Statistical analyses were per- ventilation, non-invasive ventilation and renal replace- formed with R 3.2.3 software packages (R Development ment therapy) and treatment limitation. Group com- Core Team, Vienna, Austria). parisons for continuous variables were performed using the Kruskal–Wallis test if no-normally distributed, and Results ANOVA if normally distributed; for categorical vari- This study included 1346 patients from 138 ICUs across ables the Chi square test was used. Overall survival from 28 countries. The median number of recruited patients ICU admission was estimated using the Kaplan–Meier per ICU was 7 (IQR 3–12). The study flow chart is illus- method. If lost to follow-up at 90 days, patients were cen- trated in Additional file 6. Survival at 30 and 90, respec- sored at 30 days or ICU discharge if status at 30 days was tively, was available in 97% and 90% of the cohort. Jung et al. Crit Care (2021) 25:149 Page 4 of 9 Frailty was assessed by an ICU physician in 55%, by below and above 75 years. Of note, there is no difference dedicated research staff in 21%, by an ICU nurse in 13%, between age groups in frail patients. and by other personnel in 11% of the cases. Further Survival at 30 days was 66% (63–69) in fit, 53% (47–61) information for CFS assessment was provided by hospi- in vulnerable and 41% (35–47) in frail patients. The differ- tal records (51%), the family or caregivers (25%), by the ence persisted at 90 days with respective survival of 59% patient (22%) and by other information (2%). Patients (56–63), 47% (40–55) and 33% (27–39) p < 0.001). Length characteristics and outcome were similar whether CFS of ICU stay for patients discharged alive was 15 days was rated by an ICU physician or other personal and (IQR7.0–29.0) in fit, 10 days (IQR 5–21) in vulnerable whether the CFS was based on hospital records or other and 6 days (IQR 3.0–13) in frail patients (p < 0.001). sources. Table 2 shows models revealing the association The median age of patients was 75 years (IQR 72–78, between frailty and outcome even after controlling for range 70–96. Median CFS was 3 (IQR 2–4) and 20.7% of comorbidities and disease severity and treatment strate- the patients were frail (CFS ≥ 5). Only one patient had gies. Frailty was associated with increased use of treat- CFS 9 so this group was not further split. Further base- ment limitations and reduction in respiratory support line characteristics of the study population are given in as shown in Fig. 2 (cumulative incidences in Additional Table 1. file 9). The use of treatment limitations was significantly The overall survival at 30 days was 59% (95% CI 56–62). higher in frail patients compared to fit patients and vul- Ninety-day survival was 52% and decreased with increas- nerable patients (20-day cumulative incidence was 26% ing CFS as illustrated in Additional file 7. Numbers of (95% CI 23–29) for fit patients, 40% (33–47) for vulner- deaths are reported in Additional file 8. Figure 1a shows able patients and 43% (95% CI 37–48) for frail patients the prognostic relevance of frailty in a survival analysis. (p < 0.001). The association between frailty and treat- Figure 1b illustrates the frailty category by age groups ment limitation remained statistically significant after Table 1 Patient characteristics of the study population All patients Fit (CFS 1–3) Vulnerable (CFS: 4) Frail (CFS: 5–9) n = 1346 n = 874 n = 193 n = 279 Female sex n (%) 381 (28) 209 (24) 64 (34) 108 (39) Age Median (IQR) 75 (72–78) 74 (72–77) 76 (73–79) 78 (74–82) Frailty score—CFS Median (IQR) 3 (2–4) 2 (2–3) 4 (4–4) 6 (5–7) SOFA score Median (IQR) 6 (3–8) 5 (3–8) 6 (3–8) 6 (4–9) BMI Median (IQR) 28 (25–31) 27 (25–30) 28 (25–31) 28 (24–31) PaO2/FiO2 (mmHg), n (%)   ≤ 100 485 (37) 319 (37) 65 (35) 101 (37)   > 100–200 546 (41) 376 (43) 76 (40) 94 (34)   > 200–300 173 (13) 111 (13) 25 (13) 37 (13)   > 300 125 (9) 59 (7) 11 (12) 44 (16) Comorbidities, n (%) Diabetes mellitus 471 (35) 240 (27) 89 (47) 142 (51) Ischemic heart disease 291 (22) 127 (14) 59 (31) 105 (39) Chronic renal insufficiency 211 (16) 77 (9) 31 (17) 103 (38) Arterial Hypertension 896 (67) 529 (61) 139 (73) 228 (82) Pulmonary disease 314 (24) 169 (20) 60 (32) 85 (31) Chronic heart failure 205 (15) 72 (6) 49 (26) 84 (32) CFS clinical frailty scale, SOFA sequential organ failure assessment score, IQR inter-quartile range, BMI body mass index, PaO2 partial pressure of oxygen, FiO2 fraction of inspired oxygen Jung et al. Crit Care (2021) 25:149 Page 5 of 9 Fig. 1 a Kaplan–Meier curve illustrating survival dependent on clinical frailty scale (CFS) category: fit, vulnerable and frail. b Patients were divided according to the age median (75 years) and survival was illustrated according to their frailty category adjustment for patient characteristics (aHR for frail vs fit Before the COVID-19 pandemic emerged, frailty was patients 2.79 (95% CI 1.96–3.91, p < 0.001)). already established as an important factor for outcome, A sensitivity analysis showed that the findings were particularly in very old ICU patients. This was docu- similar when excluding patients from outside Europe mented in large studies from Canada [21] and Europe, (N = 158). and also in a systematic review. As a result, frailty was suggested early on during the pandemic as a useful tool to assist guiding therapy. In the UK, NICE issued guide- Discussion lines advocating the use of the CFS in patients above Our study reveals three important findings in COVID-19 65 years to assist with decision making regarding ICU patients. First, frailty is a useful tool to stratify the risk admission. Here, scores of five and above were thought to of death at one and three months after admission to the represent a worse prognosis in critically ill patients [22]. ICU, and frailty offers an important additional prognos- However, the evidence for using frailty was extrapolated tic information to the age in patients aging 70 and older. from pre-pandemic data [23], and as a result, this guide- Second, outcome in terms of mortality in patients with a line was heavily criticised for being based on insufficient frailty level ≥ 5 is similar in patients across all age groups data. ≥ 70 years. Third, frailty was also associated with less use During this pandemic, there have been many discus- of mechanical ventilation and a higher rate of treatment sions about the care of the old and the very old critical limitation. ill patients. Frailty has been the focus in four studies. In Age is frequently associated with a higher rate of hos- a retrospective single-centre study from Italy with 105 pitalisation, ICU admission and mortality in COVID- patients, the frailty index was found to be an independ- 19 patients [18–20]. The high risk of mortality in older ent predictor not only of in-hospital mortality but also patients together with constraints on ICU bed availability for ICU admission. In another single-centre study from may raise the question of rationing ICU admissions. Age the UK of 215 hospitalised, non-ICU, patients both CFS alone should not be used and may be considered ageist and age were associated with mortality [24]. By con- [16]. For this reason, other factors should be investigated. trast, in a larger study of 1071 hospitalised patients with Frailty as assessed by the CFS is a good candidate as it COVID-19, CFS was not associated with mortality. To has previously been found to be strongly associated with date, the largest investigation of frailty in COVID-19 mortality and is easy to use in acutely ill patients. patients is a multicentre observational study from the Jung et al. Crit Care (2021) 25:149 Page 6 of 9 Table 2 Three sequential random effects and multilevel Cox regression models were used to evaluate the impact of frailty on both 30-days survival Model 1 unadjusted HR p value Model 2 adjusted HR p value Model 3 adjusted HR p value (95% CI) (95% CI) (95% CI) Survival at 30 days Vulnerable versus fit 1.75 (1.35–2.25) < 0.001 1.55 (1.14–2.10) 0.011 1.14 (0.79–1.65) 0.4811 Frail versus fit 3.20 (2.56–4.13) < 0.001 2.41 (1.77–3.27) < 0.001 1.86 (1.36–2.52) < 0.001 Treatment limitation Vulnerable versus fit 2.26 (1.73–2.96) < 0.001 1.7 (1.21–2.38) 0.0021 Frail versus fit 3.98 (3.08–5.21) < 0.001 2.79 (1.96–3.91) < 0.001 Mechanical ventilation Vulnerable versus fit 0.83 (0.67–1.01) 0.055 0.92 (0.73–1.16) 0.5 Frail versus fit 0.75 (0.62–0.92) 0.005 0.69 (0.54–0.87) 0.0043 Non-invasive ventilation Vulnerable versus fit 1.58 (1.11–2.25) 0.011 1.22 (0.79–1.88) 0.37 Frail versus fit 1.58 (1.12–2.24) 0.009 1.26 (0.8–1.95) 0.32 Non-invasive ventilation/mechanical ventilation Vulnerable versus fit 0.94 (0.78–1.13) 0.51 1 (0.81–1.24) 0.99 Frail versus fit 0.84 (0.7–1.02) 0.069 0.74 (0.58–0.91) 0.0096 Vasoactive drugs Vulnerable versus fit 0.99 (0.81–1.22) 0.93 1.09 (0.86–1.39) 0.47 Frail versus fit 1.01 (0.84–1.25) 0.88 0.9 (0.7–1.15) 0.44 Renal replacement therapy Vulnerable versus fit 1.23 (0.81–1.86) 0.33 1.14 (0.7–1.87) 0.61 Frail versus fit 1.62 (1.1–2.45) 0.014 1.01 (0.59–1.65) 0.98 First, we estimated the impact of frailty on outcome without adjustment on confounding using a baseline model including only frailty (model 1). Second, to estimate the impact of frailty when adjusting on patients’ baseline characteristics the following covariates were added to model 1: age, sex, comorbidities, SOFA score, BMI, ­PaO2/FiO2. Third, to evaluate whether the effect of frailty was independent of ICU management strategies, both organ support and treatment limitation (model 3) were added to model 2 as time-dependent covariates. For all outcomes, significance of the random centre effect was tested comparing the likelihood of two models including frailty with and without random effect. Random effect was significant for all outcomes. No violation of the proportional hazard assumption was detected in the models UK and Italy involving 1564 patients from 11 hospitals. age alone with no frailty assessment used for outcome The study included all hospitalised patients ≥ 18 years prognostication. admitted with COVID-19 during a defined period and The use of frailty in general, and CFS in particular, was therefore differs from the present study which included the focus of a recent editorial in ICM published prior to only patients ≥ 70 years. They found a large proportion the pandemic [26]. This suggested that risk stratification of patients were frail (CSF ≥ 5 in 51.4%) and that dis- should not be based on age alone but should include a ease outcome was better predicted by frailty, measured frailty assessment. They also stressed the role of putting in with the CFS, than either age or comorbidity alone. place a time limited trial of treatment on admission to ICU The importance of chronological age in COVID-19 as mortality after ICU treatment in frail elderly patients has been extensively documented. In a retrospective remains high. Current and recent research has not proven case series, 1591 consecutive patients with a median the long-term benefit of frailty assessment in these patients. age of 63 years were admitted to 72 Italian hospitals The role of frailty may therefore better inform best inter- between February and March 2020 [18]. They found est decisions—i.e. whether burdens of ICU are more likely that patients with a median age of 63 years or more to outweigh benefits (or vice versa) but caution should be had a higher mortality than younger patients; they also applied in excluding patients for ICU based on age/frailty required mechanical ventilation more frequently. This status alone as there may be some patients denied ICU that result was supported by another retrospective study still have the potential to benefit despite being frail. from Germany in 10,021 hospitalised adult patients The present research has several strengths [27]. It is a from 920 different hospitals [25]. Overall, patients aged multicentre trial that recruited patients from 28, mainly 80 years or older had the highest mortality of 72%. European, countries. It included different types of hospitals These two studies, however, focused on chronological thus reflecting diverse health care systems underlining the Jung et al. Crit Care (2021) 25:149 Page 7 of 9 Fig. 2 Cumulative incidence of organ support and treatment limitations. a Combined mechanical ventilation (MV) and non-invasive ventilation. b Mechanical ventilation (MV). c Vasoactive drugs. d Non-invasive ventilation (NIV). e Treatment limitations. f Renal replacement therapy (RRT) validity of the results. High quality data were collected pro- was not possible to assess frailty in 7% of patients due to spectively despite the strain on health care systems during insufficient information. (8) We did not record informa- the pandemic. In addition, this study focuses exclusively on tion about the ethnic background, although it might be a elderly patients who were admitted to an intensive care unit. potential confounding factor [28]. Our study, however, has a number of limitations: (1) No Our study does throw up some ethical dilemmas. There data was collected about the pre-ICU triage process and is an inter-relationship between high frailty scores and as a result we do not know how many very old critically ill the unconscious bias of health care providers. For exam- patients were denied ICU-admission. (2) it is an unblinded ple, a high CFS on admission could lead the ICU health study as it is difficult to conduct a blinded study for frailty. care provider to treat a patient less aggressively and to set (3) CFS is not a suitable tool to evaluate patients with a limitation of therapy earlier in their illness. Thus, the either temporary disability (For example as a result of knowledge of frailty implicitly influences the outcome of trauma or delirium) or stable long-term disabilities (for the patient. We raise this as a limitation of our study sim- example, cerebral palsy), learning disability or autism; in ilar to all studies describing outcomes in very old patients the inclusion and exclusion criteria of our cohort these [10, 11, 24, 29], but also as a "caveat" for future studies. considerations were not explicitly acknowledged, which On the other hand, we know that frailty is the common, may have promoted selection bias. Furthermore, there multifactorial endpoint of life, and therefore the presence are other health conditions, unrelated to frailty, that can of frailty per se (independent of its measurement) influ- limit activity that might lead to an artificially high CFS ences patient outcome and is thus a self-fulling prophecy. that does not reflect true frailty. (4) Another limitation of this study is the lack of functional outcome data. While Conclusion we were able to investigate associations with mortal- Frailty provides relevant prognostic information in ity, the extent of morbidity in survivors remains unclear. elderly COVID-19 patients in addition to age and comor- (5) No younger patients were included in this study for bidities. Therefore, we recommend that a frailty assess- comparison. (6) The list of co-morbidities recorded was ment should be routinely performed in these patients. incomplete as only the most prevalent were documented, In times of limited resources on the ICU, a frailty assess- other comorbidities such as haematologic disorders or ment of elderly patients could be included in a holistic those with immune deficiencies were not recorded. (7) It assessment of patients. Jung et al. Crit Care (2021) 25:149 Page 8 of 9 Supplementary Information Sancho, Jesús Priego, Enas M.Y. Abualqumboz, Momin Majed Yousuf Hilles, Mahmoud Saleh, Nawfel Ben-HAmouda, Andrea Roberti, Alexander Dullenkopf, The online version contains supplementary material available at https://​doi.​ Yvan Fleury, Bernardo Bollen Pinto, Joerg C. Schefold, Mohammed Al-Sadawi. org/​10.​1186/​s13054-​021-​03551-3. Hans Flaatten, Bernhard Wernly, Antonio Artigas, Michael Beil, Sigal Sviri, Peter Vernon van Heerden, Wojciech Szczeklik, Muhammed Elhadi, Tilemachos Additional file 1.: List of collaborators: COVIP-study; Description: List of Zafeiridis, Rui Moreno, Maurizio Cecconi, Ariane Boumendil. COVIP study collaborators with affiliations Authors’ contributions Additional file 2.: COVIP Country map; Distribution of study sites and All authors participated in the design and conductance of the trial and were included patients per country. The first number is the number of ICUs per involved in patient recruitment and/or study organisation. The primary country, the second the total number of included patients per country statistical analysis was performed by AB; however, several authors had access Additional file 3.: Clinical Frailty Scale; Description: Pictograms and to the data. All authors participated in the interpretation of the results. CJ, HF description of the Clinical Frailty Scale and BG wrote the initial draft of the manuscript. All other authors revised the manuscript and had access to the final version. Additional file 4.: Definition of the comorbidities; Description: Detailed definition of the comorbidities of patients included in the COVIP study Funding Additional file 5.: Definition of organ support; Description: Definition of Open Access funding enabled and organized by Projekt DEAL. The support of organ support in recruited patients to be documented in the case report the study in France by a grant from Fondation Assistance Publique-Hôpitaux form de Paris pour la recherche is greatly appreciated. In Norway, the study was supported by a grant from the Health Region West. In addition, the study Additional file 6.: Consort flow chart; Description: Consort flow chart was funded by a grant from the European Open Science Cloud (EOSC) by illustrating screening and inclusion into the COVIP study the European Commission. EOSCsecretariat.eu has received funding from the Additional file 7.: Kaplan Meier curve illustrating survival dependent European Union’s Horizon Programme call H2020-INFRAEOSC-05-2018-2019, on clinical frailty scale (CFS); Description: Kaplan Meier curve illustrating grant agreement number 831644. No further specific funding was received. survival dependent on clinical frailty scale (CFS) for each category Availability of data and materials Additional file 8.: Numbers of deaths; Description: Numbers of deaths Individual participant data that underlie the results reported in this article are reported during the study/follow-up available to investigators whose proposed use of the data has been approved Additional file 9.: Survival estimates for the primary endpoint (30-day by the COVIP steering committee. mortality) and additional time points for fit, vulnerable and frail patients as well as cumulative incidence of treatment limitations and treatment Take‑home‑message modalities Description: Table of survival estimates for the primary end- This prospective, large, international multicentre COVIP-study delineates the point (30-day mortality) and additional time points value of using frailty in the prognostication of elderly patient with COVID-19 up to three months after ICU admission. CFS is independently associated with outcome. Acknowledgements The authors want to thank all investigators and study personal for their great Tweet support of the study. The COVIP study group consists of the authors and the CFS is a valuable tool to assist prognostication in elderly ICU patients during following persons: Philipp Eller, Michael Joannidis, Dieter Mesotten, Pas- the COVID-19 pandemic. cal Reper, Sandra Oeyen, Walter Swinnen, Helene Brix, Jens Brushoej, Maja Villefrance, Helene Korvenius Nedergaard, Anders Thais Bjerregaard, Ida Riise Balleby, Kasper Andersen, Maria Aagaard Hansen, Stine Uhrenholt, Helle Declarations Bundgaard, Jesper Fjølner, Aliae AR Mohamed Hussein, Rehab Salah, Yasmin Khairy NasrEldin Mohamed Ali, Kyrillos Wassim, Yumna A. Elgazzar, Samar Ethics approval and consent to participate Tharwat, Ahmed Y. Azzam, Ayman abdelmawgoad Habib, Hazem Maarouf Ethical approval was obtained for all sites. Informed consent was obtained if Abosheaishaa, Mohammed A Azab, Susannah Leaver, Arnaud Galbois, Bertrand not waived by the local ethical approval. Guidet, Cyril Charron, Emmanuel Guerot, Guillaume Besch, Jean-Philippe Rigaud, Julien Maizel, Michel Djibré, Philippe Burtin, Pierre Garcon, Saad Consent for publication Nseir, Xavier Valette, Nica Alexandru, Nathalie Marin, Marie Vaissiere, Gaëtan Not applicable. Plantefeve, Thierry Vanderlinden, Igor Jurcisin, Buno Megarbane, Anais Cail- lard, Arnaud Valent, Marc Garnier, Sebastien Besset, Johanna Oziel, Jean-herlé Competing interests Raphaelen, Stéphane Dauger, Guillaume Dumas, Bruno Goncalves, Gaël Piton, The authors declare that they have no competing interests. JCS reports grants Christian Jung, Raphael Romano Bruno, Malte Kelm, Georg Wolff, Eberhard (full departmental disclosure) from Orion Pharma, Abbott Nutrition Interna- Barth, Ulrich Goebel, Eberhard Barth, Anselm Kunstein, Michael Schuster, Martin tional, B. Braun Medical AG, CSEM AG, Edwards Lifesciences Services GmbH, Welte, Matthias Lutz, Patrick Meybohm, Stephan Steiner, Tudor Poerner, Hendrik Kenta Biotech Ltd, Maquet Critical Care AB, Omnicare Clinical Research AG, Haake, Stefan Schaller, Detlef Kindgen-Milles, Christian Meyer, Muhammed Kurt, Nestle, Pierre Fabre Pharma AG, Pfizer, Bard Medica S.A., Abbott AG, Anandic Karl Friedrich Kuhn, Winfried Randerath, Jakob Wollborn, Zouhir Dindane, Hans- Medical Systems, Pan Gas AG Healthcare, Bracco, Hamilton Medical AG, Frese- Joachim Kabitz, Ingo Voigt, Gonxhe Shala, Andreas Faltlhauser, Nikoletta Rovina, nius Kabi, Getinge Group Maquet AG, Dräger AG, Teleflex Medical GmbH, Glaxo Zoi Aidoni, Evangelia Chrisanthopoulou, Antonios Papadogoulas, Mohan Smith Kline, Merck Sharp and Dohme AG, Eli Lilly and Company, Baxter, Astellas, Gurjar, Ata Mahmoodpoor, Abdullah khudhur Ahmed, Brian Marsh, Ahmed Astra Zeneca, CSL Behring, Novartis, Covidien, Philips Medical, Phagenesis Elsaka, Sigal Sviri, Vittoria Comellini, Ahmed Rabha, Hazem Ahmed, Silvio a Ltd, Prolong Pharmaceuticals and Nycomed outside the submitted work. The Namendys-Silva, Abdelilah Ghannam, Martijn Groenendijk, Marieke Zegers, money went into departmental funds. No personal financial gain applied. Dylan W. De Lange, Alex Cornet, Mirjam Evers, Lenneke Haas, Tom Dormans, Willem Dieperink, Luis Romundstad, Britt Sjøbø, Finn H. Andersen, Hans Frank Author details 1 Strietzel, Theresa Olasveengen, Michael Hahn, Miroslaw Czuczwar, Ryszard Department of Cardiology, Pulmonology and Vascular Medicine, Medical Gawda, Jakub Klimkiewicz, Maria de Lurdes Campos Santos, André Gordinho, Faculty, Heinrich-Heine-University Duesseldorf, Moorenstraße 5, 40225 Dues- Henrique Santos, Rui Assis, Ana Isabel Pinho Oliveira, Mohamed Raafat Badawy, seldorf, Germany. 2 Department of Clinical Medicine, University of Bergen, David Perez-Torres, Gemma Gomà, Mercedes Ibarz Villamayor, Angela Prado Bergen, Norway. 3 Department of Anaesthesia and Intensive Care, Haukeland Mira, Patricia Jimeno Cubero, Susana Arias Rivera, Teresa Tomasa, David Iglesias, University Hospital, Bergen, Norway. 4 Department of Intensive Care, Aarhus Eric Mayor Vázquez, Cesar Aldecoa, Aida Fernández Ferreira, Begoña Zalba- University Hospital, Aarhus, Denmark. 5 Department of Cardiology, Paracelsus Etayo, Isabel Canas-Perez, Luis Tamayo-Lomas, Cristina Diaz-Rodriguez, Susana Medical University, Salzburg, Austria. 6 Department of Intensive Care Medicine, Jung et al. Crit Care (2021) 25:149 Page 9 of 9 CIBER Enfermedades Respiratorias, Corporacion Sanitaria Universitaria Parc Tauli, 12. Owen RK, Conroy SP, Taub N, Jones W, Bryden D, Pareek M, Faull C, Autonomous University of Barcelona, Sabadell, Spain. 7 Department of Acute Abrams KR, Davis D, Banerjee J. Comparing associations between frailty Medicine, Geneva University Hospitals, Geneva, Switzerland. 8 Department and mortality in hospitalised older adults with or without COVID-19 of Intensive Care Medicine, Inselspital, Universitätsspital, University of Bern, infection: a retrospective observational study using electronic health Bern, Switzerland. 9 Department of Medical Intensive Care, Hadassah University records. Age Ageing. 2021;50(2):307–16. Medical Center, Jerusalem, Israel. 10 General Intensive Care Unit, Hadas- 13. Woolford SJ, D’Angelo S, Curtis EM, Parsons CM, Ward KA, Dennison EM, sah University Medical Center, Jerusalem, Israel. 11 Center for Intensive Care Patel HP, Cooper C, Harvey NC. COVID-19 and associations with frailty and and Perioperative Medicine, Jagiellonian University Medical College, Krakow, multimorbidity: a prospective analysis of UK Biobank participants. Aging Poland. 12 2nd Department of Anesthesiology and Intensive Care, Medical Clin Exp Res. 2020;32(9):1897–905. University of Lublin, Staszica 16, 20‑081 Lublin, Poland. 13 Faculty of Medicine, 14. Darvall JN, Bellomo R, Young PJ, Rockwood K, Pilcher D. Frailty and mor- University of Tripoli, Tripoli, Libya. 14 Division of Intensive Care and Emergency tality in patients with COVID-19. Lancet Public Health. 2020;5(11):e580. Medicine, Department of Internal Medicine, Medical University Innsbruck, 15. Rockwood K, Song X, MacKnight C, Bergman H, Hogan DB, McDowell Innsbruck, Austria. 15 Department of Intensive Care 1K12IC, Ghent University I, Mitnitski A. A global clinical measure of fitness and frailty in elderly Hospital, Ghent, Belgium. 16 Intensive Care Unit General Hospital of Larissa, people. CMAJ. 2005;173(5):489–95. Larissa, Greece. 17 Mater Misericordiae University Hospital, Dublin, Ireland. 16. Guidet B, Flaatten H, Boumendil A, Morandi A, Andersen FH, Artigas A, 18 Department of Anaesthesia and Intensive Care, Ålesund Hospital, Ålesund, Bertolini G, Cecconi M, Christensen S, Faraldi L, et al. Withholding or with- Norway. 19 Department of Circulation and Medical Imaging, Norwegian Univer- drawing of life-sustaining therapy in older adults (>/= 80 years) admitted sity of Science and Technology, Trondheim, Norway. 20 Unidade de Cuidados to the intensive care unit. Intensive Care Med. 2018;44(7):1027–38. Intensivos Neurocríticos e Trauma, Hospital de São José, Centro Hospitalar 17. Guidet B, Hodgson E, Feldman C, Paruk F, Lipman J, Koh Y, Vincent Universitário de Lisboa Central, Faculdade de Ciências Médicas de Lisboa, Nova JL, Azoulay E, Sprung CL. The Durban World Congress Ethics Round Médical School, Lisbon, Portugal. 21 Department of Anaesthesia, IRCCS Instituto Table Conference Report: II. Withholding or withdrawing of treatment Clínico Humanitas, Humanitas University, Milan, Italy. 22 General Intensive Care, in elderly patients admitted to the intensive care unit. J Crit Care. St George’s University Hospitals NHS Foundation Trust, London, UK. 23 Sor- 2014;29(6):896–901. bonne Universités, UPMC Univ Paris 06, INSERM, UMR_S 1136, Institut Pierre 18. Grasselli G, Greco M, Zanella A, Albano G, Antonelli M, Bellani G, Louis d’Epidémiologie et de Santé Publique, Equipe: épidémiologie hospitalière Bonanomi E, Cabrini L, Carlesso E, Castelli G, et al. Risk factors associated qualité et organisation des soins, 75012 Paris, France. 24 Assistance Publique with mortality among patients with COVID-19 in intensive care units in - Hôpitaux de Paris, Hôpital Saint-Antoine, service de réanimation médicale, Lombardy. Italy. JAMA Intern Med. 2020;180(10):1345–55. 75012 Paris, France. 25 Department of Intensive Care Medicine, University Medi- 19. Gupta S, Hayek SS, Wang W, Chan L, Mathews KS, Melamed ML, Brenner cal Center, University Utrecht, Utrecht, The Netherlands. SK, Leonberg-Yoo A, Schenck EJ, Radbel J, et al. Factors associated with death in critically ill patients with coronavirus disease 2019 in the US. Received: 17 February 2021 Accepted: 25 March 2021 JAMA Intern Med. 2020;180(11):1–12. 20. Network C-IGobotR, the C-ICUI: Clinical characteristics and day-90 outcomes of 4244 critically ill adults with COVID-19: a prospective cohort study. Intensive Care Med. 2021;47(1):60–73 21. Heyland D, Cook D, Bagshaw SM, Garland A, Stelfox HT, Mehta S, Dodek References P, Kutsogiannis J, Burns K, Muscedere J, et al. The very elderly admitted to 1. Maltese G, Corsonello A, Di Rosa M, Soraci L, Vitale C, Corica F, Lattan- ICU: a quality finish? Crit Care Med. 2015;43(7):1352–60. zio F. Frailty and COVID-19: a systematic scoping review. J Clin Med. 22. COVID-19 rapid guideline: critical care in adults. 2020;9(7):66. 23. Azoulay E, Beloucif S, Guidet B, Pateron D, Vivien B, Le Dorze M. Admission 2. Alkuzweny M, Raj A, Mehta S. Preparing for a COVID-19 surge: ICUs. EClini- decisions to intensive care units in the context of the major COVID-19 calMedicine. 2020;25:100502. outbreak: local guidance from the COVID-19 Paris-region area. Crit Care. 3. Flaatten H, De Lange DW, Morandi A, Andersen FH, Artigas A, Bertolini 2020;24(1):293. G, Boumendil A, Cecconi M, Christensen S, Faraldi L, et al. The impact of 24. Chinnadurai R, Ogedengbe O, Agarwal P, Money-Coomes S, Abdurrah- frailty on ICU and 30-day mortality and the level of care in very elderly man AZ, Mohammed S, Kalra PA, Rothwell N, Pradhan S. Older age and patients (>/= 80 years). Intensive Care Med. 2017;43(12):1820–8. frailty are the chief predictors of mortality in COVID-19 patients admitted 4. Guidet B, de Lange DW, Boumendil A, Leaver S, Watson X, Boulanger C, to an acute medical unit in a secondary care setting—a cohort study. Szczeklik W, Artigas A, Morandi A, Andersen F, et al. The contribution of BMC Geriatr. 2020;20(1):409. frailty, cognition, activity of daily life and comorbidities on outcome in 25. Karagiannidis C, Mostert C, Hentschker C, Voshaar T, Malzahn J, Schillinger acutely admitted patients over 80 years in European ICUs: the VIP2 study. G, Klauber J, Janssens U, Marx G, Weber-Carstens S, et al. Case charac- Intensive Care Med. 2020;46(1):57–69. teristics, resource use, and outcomes of 10 021 patients with COVID-19 5. Muscedere J, Waters B, Varambally A, Bagshaw SM, Boyd JG, Maslove admitted to 920 German hospitals: an observational study. Lancet Respir D, Sibley S, Rockwood K. The impact of frailty on intensive care unit Med. 2020;8(9):853–62. outcomes: a systematic review and meta-analysis. Intensive Care Med. 26. Ranzani OT, Besen B, Herridge MS. Focus on the frail and elderly: 2017;43(8):1105–22. Who should have a trial of ICU treatment? Intensive Care Med. 6. Flaatten H, Guidet B, Andersen FH, Artigas A, Cecconi M, Boumendil A, 2020;46(5):1030–2. Elhadi M, Fjolner J, Joannidis M, Jung C, et al. Reliability of the Clinical 27. Alexander PE, Debono VB, Mammen MJ, Iorio A, Aryal K, Deng D, Brocard Frailty Scale in very elderly ICU patients: a prospective European study. E, Alhazzani W. COVID-19 coronavirus research has overall low methodo- Ann Intensive Care. 2021;11(1):22. logical quality thus far: case in point for chloroquine/hydroxychloroquine. 7. Lewis EG, Breckons M, Lee RP, Dotchin C, Walker R. Rationing care by J Clin Epidemiol. 2020;123:120–6. frailty during the COVID-19 pandemic. Age Ageing. 2021;50(1):7–10. 28. Apea VJ, Wan YI, Dhairyawan R, Puthucheary ZA, Pearse RM, Orkin CM, Prowle 8. Aw D, Woodrow L, Ogliari G, Harwood R. Association of frailty with mortality in JR. Ethnicity and outcomes in patients hospitalised with COVID-19 infection in older inpatients with COVID-19: a cohort study. Age Ageing. 2020;49(6):915–22. East London: an observational cohort study. BMJ Open. 2021;11(1):e042140. 9. Labenz C, Schattenberg JM, Worns MA, Sprinzl MF. Response: frailty 29. Bellelli G, Rebora P, Citerio G. The role of frailty in COVID-19 patients. assessment in the COVID-19 pandemic. J Investig Med. 2020;68(7):1302. Intensive Care Med. 2020;46(10):1958–9. 10. Hewitt J, Carter B, Vilches-Moraga A, Quinn TJ, Braude P, Verduri A, Pearce L, Stechman M, Short R, Price A, et al. The effect of frailty on survival in patients with COVID-19 (COPE): a multicentre, European, observational Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in pub- cohort study. Lancet Public Health. 2020;5(8):e444–51. lished maps and institutional affiliations. 11. Bellelli G, Rebora P, Valsecchi MG, Bonfanti P, Citerio G. members C-MT: frailty index predicts poor outcome in COVID-19 patients. Intensive Care Med. 2020;46(8):1634–6.