Articles Prognostic implications of comorbidity patterns in critically ill COVID-19 patients: A multicenter, observational study n D. Benítez,a,b,1 Jordi de Batlle,a,b,1 Gerard Torres,a,b Jessica Gonza Iva lez,a,b David de Gonzalo-Calvo,a,b Adriano D.S. Targa,a,b a,b a,b  Clara Gort-Paniello, Anna Moncusí-Moix, Adrian Ceccato, Laia Ferna b,c ndez-Barat,b,d Ricard Ferrer,b,e Dario Garcia-Gasulla,f Rosario Menendez,b,g Anna Motos,b,d Oscar Pe~ nuelas,b,h Jordi Riera,b,e Jes us F. Bermejo-Martin,b,i Yhivian Pe~ nasco,j Pilar Ricart,k María Cruz Martin Delgado, Luciano Aguilera, Alejandro Rodríguez, Maria Victoria Boado Varela, Fernando Suarez-Sipmann,p l m n o Juan Carlos Pozo-Laderas,q Jordi Sole-Violan,r Maite Nieto,s Mariana Andrea Novo,t Jose Barbera n,u Rosario Amaya Villar,v w x Jose Garnacho-Montero, Jose Luis García-Garmendia, Jose M. G  omez, Jose Angel Lorente, Aaron Blandino Ortiz,z y b,h Luis Tamayo Lomas,aa Esther Lo pez-Ramos,ab Alejandro Ubeda,  ac Mercedes Catala n-Gonzalez,ad Angel Sa nchez-Miralles,ae Ignacio Martínez Varela, Ruth Noemí Jorge García, Nieves Franco, Víctor D. Gumucio-Sanguino, Arturo Huerta Garcia,aj af ag ah ai Elena Bustamante-Munguira,ak Luis Jorge Valdivia,al Jes us Caballero,am Elena Gallego,an Amalia Martínez de la Ga ndara,ao  Alvaro Castellanos-Ortega, Josep Trenado, Judith Marin-Corral, Guillermo M Albaiceta, Maria del Carmen de la Torre,at ap aq ar b,as Ana Loza-Va zquez,au Pablo Vidal,av Juan Lopez Messa,aw Jose M. An ~on,b,ax Cristina Carbajales Perez,ay Victor Sagredo,az ba bb bc Neus Bofill, Nieves Carbonell, Lorenzo Socias, Carme Barbera , Angel Estella,be Manuel Valledor Mendez,bf Emili Diaz,bg bd Ana L opez Lago, Antoni Torres, and Ferran Barbe *, on behalf of the CIBERESUCICOVID Project (COV20/00110, ISCIII)2 bh b,d a,b a Translational Research in Respiratory Medicine, University Hospital Arnau de Vilanova and Santa Maria, IRBLleida, Lleida, Spain b CIBER of Respiratory Diseases (CIBERES), Institute of Health Carlos III, Madrid, Spain c  i Innovacio Critical Care Center, ParcTaulí Hospital Universitari, Institut d'Investigacio  Parc Taulí I3PT, Sabadell, Spain d Department of Pneumology, Hospital Clinic of Barcelona; August Pi i Sunyer Biomedical Research Institute−IDIBAPS, University of Barcelona, Barcelona, Spain e Intensive Care Department, Vall d’Hebron Hospital Universitari. SODIR Research Group, Vall d’Hebron Institut de Recerca (VHIR), Barcelona, Spain f Barcelona Supercomputing Center (BSC), Barcelona, Spain g Pulmonology Service, University and Polytechnic Hospital La Fe, Valencia, Spain h Hospital Universitario de Getafe, Madrid, Spain; Universidad Europea, Madrid, Spain i Hospital Universitario Río Hortega de Valladolid, Valladolid, Spain; Group for Biomedical Research in Sepsis (BioSepsis), Instituto de Investigacion Biom edica de Salamanca (IBSAL), Salamanca, Spain j Servicio de Medicina Intensiva, Hospital Universitario Marqu es de Valdecilla, Santander, Spain k Servei de Medicina Intensiva, Hospital Universitari Germans Trias, Badalona, Spain l Hospital Universitario Torrejon- Universidad Francisco de Vitoria, Madrid, Spain m Servicio de Anestesiología y Reanimaci on, Hospital Universitario Basurto, Bilbao, Spain n Critical Care Department, Hospital Joan XXIII, Tarragona, Spain o Servicio de Medicina Intensiva, Hospital de Cruces, Baracaldo, Vizcaya, Spain p Intensive Care Unit, Hospital Universitario La Princesa, Madrid, Spain q UGC-Medicina Intensiva, Hospital Universitario Reina Sofia, Instituto Maimonides IMIBIC, C ordoba, Spain r Critical Care Department, Hospital Dr. Negrín Gran Canaria, Las Palmas, Gran Canaria, Spain. Universidad Fernando Pessoa, Canarias, Spain s Hospital Universitario de Segovia, Segovia, Spain t Servei de Medicina Intensiva, Hospital Universitari Son Espases, Palma de Mallorca, Illes Balears, Spain u Hospital Universitario HM Montepríncipe, Universidad San Pablo-CEU, Madrid, Spain v Intensive Care Clinical Unit, Hospital Universitario Virgen de Rocío, Sevilla, Spain w Intensive Care Clinical Unit, Hospital Universitario Virgen Macarena, Seville, Spain x Intensive Care Unit, Hospital San Juan de Dios del Aljarafe, Bormujos, Sevilla, Spain y Hospital General Universitario Gregorio Maran ~o n, Madrid, Spain z Servicio de Medicina Intensiva, Hospital Universitario Ramo n y Cajal, Madrid, Spain aa Critical Care Department, Hospital Universitario Río Hortega de Valladolid, Valladolid, Spain *Corresponding author at: University Hospital Arnau de Vilanova and Santa Maria, Translational Research in Respiratory Medi- cine, IRBLleida, Avda Alcalde Rovira Roure 80, 25198 Lleida, Spain. E-mail address: febarbe.lleida.ics@gencat.cat (F. Barbe). 1 Ivan D Benıtez and Jordi de Batlle contributed equally to this manuscript. 2 CIBERESUCICOVID Project (COV20/00110, ISCIII) collaborative group end of the document. www.thelancet.com Vol 18 Month July, 2022 1 Articles ab Servicio de Medicina Intensiva, Hospital Universitario Príncipe de Asturias, Madrid, Spain ac Servicio de Medicina Intensiva, Hospital Punta de Europa, Algeciras, Spain ad Department of Intensive Care Medicine, Hospital Universitario 12 de Octubre, Madrid, Spain ae Hospital de Sant Joan d’Alacant, Alacant, Spain af Critical Care Department, Hospital Universitario Lucus Augusti, Lugo, Spain ag Intensive Care Department, Hospital Nuestra Sen ~ora de Gracia, Zaragoza, Spain ah Hospital Universitario de Mo stoles, Madrid, Spain ai Department of Intensive Care. Hospital Universitari de Bellvitge, L’Hospitalet de Llobregat, Barcelona, Spain. Bellvitge Biomedical Research Institute (IDIBELL), L'Hospitalet de Llobregat, Barcelona, Spain aj Pulmonary and Critical Care Division; Emergency Department, Clínica Sagrada Família, Barcelona, Spain ak Department of Intensive Care Medicine, Hospital Clínico Universitario Valladolid, Valladolid, Spain al n, Leo Hospital Universitario de Leo n, Spain am Critical Care Department, Hospital Universitari Arnau de Vilanova; IRBLleida, Lleida, Spain an Unidad de Cuidados Intensivos, Hospital Universitario San Pedro de Alcantara, Caceres, Spain ao Department of Intensive Medicine, Hospital Universitario Infanta Leonor, Madrid, Spain ap Servicio de medicina intensiva. Hospital Universitario y Politecnico La Fe, Valencia, Spain aq Servicio de Medicina Intensiva, Hospital Universitario M utua de Terrassa, Terrassa, Barcelona, Spain ar Critical Care Department, Hospital del Mar-IMIM, Barcelona, Spain as Departamento de Biología Funcional. Instituto Universitario de Oncología del Principado de Asturias, Universidad de Oviedo; Instituto de Investigaci on Sanitaria del Principado de Asturias, Hospital Central de Asturias, Oviedo, Spain at Hospital de Mataro  de Barcelona, Spain au Unidad de Medicina Intensiva, Hospital Universitario Virgen de Valme, Sevilla, Spain av Complexo Hospitalario Universitario de Ourense, Ourense, Spain aw Complejo Asistencial Universitario de Palencia, Palencia, Spain ax Servicio de Medicina Intensiva. Hospital Universitario La Paz, IdiPAZ, Madrid, Spain ay  Intensive Care Unit, Hospital Alvaro Cunqueiro, Vigo, Spain az Hospital Universitario de Salamanca, Salamanca, Spain ba Department of Physical Medicine and Rehabilitation, Hospital Verge de la Cinta, Tortosa, Tarragona, Spain bb Intensive Care Unit, Hospital Clínico y Universitario de Valencia, Valencia, Spain bc Intensive Care Unit, Hospital Son Llatzer, Palma de Mallorca, Illes Balears, Spain bd Hospital Santa Maria; IRBLleida, Lleida, Spain be Intensive Care Unit, University Hospital of Jerez. Medicine Department University of Cadiz. INiBICA, Spain bf Hospital Universitario San Agustín, Asturias, Spain bg Department of Medicine, Universitat Auto noma de Barcelona (UAB); Critical Care Department, Corporacio  Sanitaria Parc Taulí, Sabadell, Barcelona, Spain bh Department of Intensive care Medicine, Complejo Hospitalario Universitario de Santiago de Compostela, Santiago de Compostela, Spain Summary The Lancet Regional Background The clinical heterogeneity of COVID-19 suggests the existence of different phenotypes with prognostic Health - Europe implications. We aimed to analyze comorbidity patterns in critically ill COVID-19 patients and assess their impact 2022;18: 100422 on in-hospital outcomes, response to treatment and sequelae. Published online xxx https://doi.org/10.1016/j. lanepe.2022.100422 Methods Multicenter prospective/retrospective observational study in intensive care units of 55 Spanish hospitals. 5866 PCR-confirmed COVID-19 patients had comorbidities recorded at hospital admission; clinical and biological parameters, in-hospital procedures and complications throughout the stay; and, clinical complications, persistent symptoms and sequelae at 3 and 6 months. Findings Latent class analysis identified 3 phenotypes using training and test subcohorts: low-morbidity (n=3385; 58%), younger and with few comorbidities; high-morbidity (n=2074; 35%), with high comorbid burden; and renal- morbidity (n=407; 7%), with chronic kidney disease (CKD), high comorbidity burden and the worst oxygenation pro- file. Renal-morbidity and high-morbidity had more in-hospital complications and higher mortality risk than low- morbidity (adjusted HR (95% CI): 1.57 (1.34-1.84) and 1.16 (1.05-1.28), respectively). Corticosteroids, but not tocilizu- mab, were associated with lower mortality risk (HR (95% CI) 0.76 (0.63-0.93)), especially in renal-morbidity and high-morbidity. Renal-morbidity and high-morbidity showed the worst lung function throughout the follow-up, with renal-morbidity having the highest risk of infectious complications (6%), emergency visits (29%) or hospital readmissions (14%) at 6 months (p<0.01). 2 www.thelancet.com Vol 18 Month July, 2022 Articles Interpretation Comorbidity-based phenotypes were identified and associated with different expression of in-hospital complications, mortality, treatment response, and sequelae, with CKD playing a major role. This could help clini- cians in day-to-day decision making including the management of post-discharge COVID-19 sequelae. Funding ISCIII, UNESPA, CIBERES, FEDER, ESF. Copyright Ó 2022 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/) Keywords: COVID-19; Critical Care; Prognosis implications independently of patient’s age, and doing Research in context this could help clinicians in day-to-day decision making including the management of post-discharge sequelae. Evidence before this study Our research presents a straightforward means of phe- notyping patients based solely on previous comorbid- COVID-19 has a broad range of manifestations, from fully ities, and provides data on what to expect from each asymptomatic to a severe life-threatening illness requiring group of patients in terms of in-hospital complications admission to the intensive care unit (ICU). Early studies and mortality, response to standard treatment, and the have identified the individual prognostic value of older prevalence of sequelae and their progressive remission age, male sex, and several chronic conditions, including up to six months after discharge. obesity, hypertension, diabetes, chronic obstructive pulmo- nary disease, cancer, chronic kidney disease (CKD) and immunosuppression status. Stepping further, some studies have tried to identify COVID-19 phenotypes based on com- plex sets of data including sociodemographic variables, baseline comorbidities, signs and symptoms during the Introduction acute phase, functional and biological parameters and COVID-19 has a broad range of manifestations, from chest radiological features. However, these studies have fully asymptomatic to a severe life-threatening illness two key drawbacks: (i) they have a limited sample size or are based on too complex sets of data not easily available requiring admission to the intensive care unit (ICU).1,2 to clinicians worldwide, which limits generalizability; (ii) Great efforts have been devoted to characterizing severe they do not assess the impact of COVID-19 phenotypes on COVID-19 patients and identifying key prognostic varia- prognosis, response to treatment and sequelae, which is bles. In this regard, early studies have identified the required for a truly comprehensive assessment of the rele- prognostic value of older age, male sex, and several vance of the identified phenotypes. chronic conditions, including obesity, hypertension, dia- betes, chronic obstructive pulmonary disease (COPD), Added value of this study cancer, chronic kidney disease (CKD), and immunosup- The present study differs from the aforementioned in pression status.3−5 The heterogeneity of comorbidity being based exclusively on the baseline comorbidity pat- patterns in such patients could be related to different terns of a large multicenter sample of critically ill patients clinical phenotypes that could show distinct prognoses, and assessing how different comorbidity backgrounds responses to treatment and sequelae. Therefore, the influence clinical outcome, response to available treatment, results of an in depth analysis of comorbidity patterns risk of different complications, mortality, and sequelae. As could have implications for clinical practice. expected, two of the identified phenotypes corresponded To date, several studies have aimed to identify phe- to a low-comorbidity and a high-comorbidity group of notypes among hospitalized COVID-19 patients based patients, which is no news for clinicians. However, we found a third phenotype characterized by the presence of on sociodemographic variables, baseline comorbidities, CKD, which had the worst prognosis at all stages (through- signs and symptoms during the acute phase, and the out hospitalization and when considering post-discharge results of complementary tests, such as biological sequelae). Moreover, even amongst patients in the low- parameters in the blood test6−9 or chest radiological comorbidity phenotype, having CKD was associated with features.8,10 Most interestingly, several studies have in-hospital mortality, which confirms the key role of CKD focused on patients admitted to the ICU that are those on COVID-19 prognosis. All these findings were indepen- at a higher risk of experiencing severe complications dent of patient’s age. and sequelae.11−17 However, no comprehensive study has yet identified the comorbidity patterns of critically Implications of all the available evidence ill COVID-19 patients and assessed its independent Critically ill COVID-19 patients can be grouped into dif- impact on prognosis, response to treatment and ferent comorbidity-based phenotypes with prognosis sequelae. www.thelancet.com Vol 18 Month July, 2022 3 Articles Therefore, the current study aims to use data from characteristics of the hospital stay (length of stay, length the CIBERESUCICOVID study18 to analyze comorbidity of ICU, and mortality). Finally, post-discharge data patterns in critically ill COVID-19 patients and assess included persistent symptoms (fatigue, cardiac compli- their impact on in-hospital parameters and outcomes, cations, and infectious complications), hospital read- response to treatment and sequelae. missions (emergency visits and hospital admissions), late clinical complications (infections, thrombotic events, atrial fibrillation and heart failure), sequelae at 3 Methods and 6 months, assessed by a thoracic CT scan (radiolog- ical normalization, persistent infiltrates, Interstitial Study design lung disease, pulmonary embolism, fibrous tracts, CIBERESUCICOVID is a multicenter prospective/retro- emphysema and other alterations), functional respira- spective observational study of critically ill COVID-19 tory test (FVC, FEV1, FEV1 to FVC ratio and DLCO) and patients admitted to the ICUs of 55 Spanish hospitals quality-of-life questionnaire. and registered in ClinicalTrials.gov with identifier NCT04457505. CIBERESUCICOVID started in May 2020 by collecting retrospective data of patients admit- Primary outcome ted to participating ICUs before May 2020 and contin- In-hospital mortality was considered as the primary out- ued prospectively from then onward (data collection is come, including all causes of death. Time to event was still ongoing). The data for the current analyses corre- calculated from ICU admission to death or discharge. spond to consecutive COVID-19 patients admitted to 55 This was assessed primarily according to comorbidity Spanish ICUs from February 2020 to December 2021. phenotypes, but also according to treatment, multimor- bidity subphenotypes, and individual comorbidities in a low-morbidity sub-population. Study population All included patients were admitted to the ICU due to the severity of COVID-19. COVID-19 diagnosis was con- Secondary outcomes firmed by a positive nasopharyngeal swab polymerase The following secondary outcomes were considered: in- chain reaction (PCR) test for SARS-CoV-2. Patients lack- hospital complications; hospital re-admissions at 3 and ing baseline or discharge data were excluded from the 6 months after discharge; and complications and current analyses. Patients transferred to other hospitals sequelae at 3 and 6 months after discharge. during or after ICU admission, receiving palliative care, or with severe mental disability precluding pulmonary function tests after discharge, were excluded from the Ethics and data protection follow-up. Given the nature of our study and the tar- The study was designed and conducted in compliance geted participants, no Patient and Public Involvement with the Declaration of Helsinki and national and inter- was possible. national law on data protection. Each of the participat- ing hospitals obtained approval from their ethics boards. Informed consent was obtained from all partici- Measures pants or their relatives when possible, and in cases Baseline variables were collected at hospital admission when this was unfeasible, an informed consent waiver and included sociodemographic, anthropometric and was authorized by the ethics board. All data were pseu- lifestyle variables as well as comorbidities registered in donymized and stored in a REDCap database hosted in electronic medical records (see “Methods. Collection of the Centro de Investigaci on Biomedica en Red (CIBER), chronic conditions” in the online supplement). In addi- Madrid, Spain. The study coordinators ensured integrity tion, the following variables were collected at the time and timely completion of data collection. of ICU admission and throughout the ICU stay: clinical (vital signs and symptoms) and biological parameters (blood test and blood gas test), including estimations of Identification of comorbidity patterns the glomerular filtration rate obtained using the 2021 Potentially relevant comorbidities (n=17) were included Chronic Kidney Disease Epidemiology Collaboration to identify morbidity patterns within the CIBERESUCI- creatinine equation19; procedures performed before and COVID population. To identify morbidity patters and during the ICU stay, including use of invasive mechani- evaluate its reproducibility, the study cohort was divided cal ventilation (IMV), hemodialysis or hemofiltration, (ratio 1:1) into training (n = 2933) and test (n = 2933) and pharmacological treatment (inotropes/vasopres- subcohorts using simple random sampling. Morbidity sors, antivirals); in-hospital complications such as infec- patterns were first identified in the training cohort tions (coinfections or nosocomial infections), using Latent class analysis (LCA).20 The number of thrombotic events, heart failure, myocarditis/pericardi- latent classes was determined using the parsimony cri- tis, delirium, shock and hemorrhages; and teria based on the minimum value of the Bayesian 4 www.thelancet.com Vol 18 Month July, 2022 Articles information criterion measure from 0 to 10 latent clas- Odds ratios and confidence intervals were graphically ses. To assess the reproducibility of the identified latent represented with a forest plot. Hospital complications classes, an independent latent class analysis was applied were also assessed according to multimorbidity pat- to the test cohort fixing the number of classes to the terns. Linear or logistic regression models were used as number obtained in the training cohort. Latent class appropriate to assess the risk of sequelae, taking into identified in both cohorts (training and test) were com- account confounding factors associated with lung dam- pared. Finally, we carried out a latent class analysis on age prior to SARS-CoV-2 infection (age, sex, chronic the whole CIBERESUCICOVID cohort (training and lung disease and asthma). Additionally, since one of the test altogether). Average posterior probabilities above latent classes grouped patients with low morbidity bur- 70% were considered as an optimal fit.21 Each patient den, the impact of individual comorbidities on in-hospi- was assigned to one class according to his or her highest tal mortality was evaluated in this population using Cox computed probability of membership. The Global model adjusted for confounding factors such as age and cohort was used to evaluate primary and secondary out- sex. comes. Prevalence of comorbidities and clinical data at Missing values were not imputed and models ICU admission were graphically represented for each included complete cases. Tables and figures present the cluster. The continuous variables of clinical data at ICU number of evaluable patients in each comparison. The admission were standardized. Given that a latent class main findings of the previous analyzes were repre- grouped patients with a high morbidity burden, it was sented in graphical abstract. decided to apply a latent class analysis to this population R statistical software, version 4.0.1 (R Project for Sta- to identify patterns of multimorbidity. tistical Computing), was used for all analyses. Statistical analyses Role of the funding source All descriptive and inferential analyses were performed Funding sources had no role in the study’s design, con- using the whole CIBERESUCICOVID cohort and based duct, and reporting. on the phenotypes obtained using the whole cohort. Descriptive statistics were used to summarize the char- acteristics of the study population. Absolute and relative Results frequencies were used for qualitative data. The means (sd) and medians (25th−75th percentile) were estimated Phenotypes based on comorbidities for quantitative variables with normal and non-normal Three statistically and clinically significant comorbidity distributions, respectively. Normal distributions were phenotypes were identified in the training cohort assessed by the Shapiro−Wilk test. Clinical data, at UCI (n = 2933) and then validated in the test cohort admission, were compared between phenotypes using (n = 2933) (Table 1 and eFigure 1). The baseline charac- ANOVA (or Kruskall-Wallis test for variables with non- teristics were similar between subcohorts (eTable1). The normal distribution) for continuous variables and chi- final classification of patients was performed by LCA on squared test (or Fisher-Freeman-Halton exact test when the global cohort (Figure 1A). LCA model showed an the expected frequencies were less than 5 in some cell) optimal fit with a mean posterior probabilities of class for qualitative. membership of 84%.21 The first phenotype, named low- Primary outcome was defined as in-hospital mortal- morbidity, included 3385 (57.7%) patients characterized ity. The effect of comorbidity phenotypes on the risk of by a low comorbidity burden (median [p25; p75] of 1 [0; 2] in-hospital mortality was assessed using Cox model comorbidity), with obesity (29.9%) as the most preva- adjusted for confounding factors such as age and sex. lent comorbidity and a younger age than the other phe- Additionally, a Fine-Gray’s competing risk model was notypes. The second phenotype, high-morbidity, performed to control for potential overestimation of in- included 2074 (35.3%) patients with a high comorbidity hospital mortality risk when considering discharge as burden (median of 3 [3; 4] comorbidities), with hyper- censored information.22,23 The same analysis were per- tension (87.6%), diabetes (52.2%), other metabolic dis- form in multimorbidity phenotypes. Furthermore, the orders (47.3%) and obesity (44.8%) being the most impact of the most broadly used pharmacological treat- frequent comorbidities. The third phenotype, named ment for COVID-19, tocilizumab and corticosteroids, renal-morbidity, included 407 (6.9%) patients with pre- on in-hospital mortality was evaluated according to vious chronic kidney disease and a very high comorbid- comorbidity phenotypes, including phenotype and drug ity burden (median of 5 [4; 6] comorbidities). Table 2 interaction terms. and Figure 1A show the main characteristics of the phe- The odds of having hospital complications were notypes at hospital admission. Reported Symptoms assessed for each comorbidity and phenotype using before hospital admission were similar in all pheno- logistic regression models adjusted for age and sex, types (eFigure 2). A Flowchart of the study including with a low-morbidity phenotype as a reference. The the initial number of evaluable patients, patients www.thelancet.com Vol 18 Month July, 2022 5 Articles Training cohort Test cohort Low-morbidity High-morbidity Renal-morbidity Low-morbidity High-morbidity Renal-morbidity n = 1488 n(%) n = 1256 n(%) n = 189 n(%) n = 1729 n(%) n = 997 n(%) n = 207 n(%) Comorbidities Obesity 368 (24.7%) 619 (49.3%) 66 (34.9%) 520 (30.1%) 441 (44.2%) 71 (34.3%) Hypertension 237 (15.9%) 1074 (85.5%) 174 (92.1%) 381 (22.0%) 892 (89.5%) 197 (95.2%) Diabetes mellitus (Type I / II) 51 (3.43%) 574 (45.7%) 99 (52.4%) 92 (5.32%) 536 (53.8%) 106 (51.2%) Chronic heart disease 14 (0.94%) 310 (24.7%) 70 (37.0%) 49 (2.83%) 241 (24.2%) 69 (33.3%) Chronic renal disease 10 (0.67%) 0 (0.00%) 186 (98.4%) 18 (1.04%) 10 (1.00%) 199 (96.1%) Chronic moderate liver disease 8 (0.54%) 12 (0.96%) 4 (2.12%) 11 (0.64%) 15 (1.50%) 8 (3.86%) Chronic mild liver disease 28 (1.88%) 39 (3.11%) 10 (5.29%) 16 (0.93%) 39 (3.91%) 11 (5.31%) Chronic neurological disease 36 (2.42%) 111 (8.84%) 17 (8.99%) 69 (3.99%) 71 (7.12%) 17 (8.21%) Chronic pulmonary disease 55 (3.70%) 203 (16.2%) 47 (24.9%) 69 (3.99%) 167 (16.8%) 41 (19.8%) Asthma 109 (7.33%) 63 (5.02%) 11 (5.82%) 109 (6.30%) 62 (6.22%) 7 (3.38%) Dementia 7 (0.47%) 10 (0.80%) 4 (2.12%) 5 (0.29%) 13 (1.30%) 1 (0.48%) Rheumatic disease 44 (2.96%) 90 (7.17%) 9 (4.76%) 44 (2.54%) 58 (5.82%) 20 (9.66%) Gastrointestinal/pancreatic 68 (4.57%) 122 (9.71%) 26 (13.8%) 83 (4.80%) 96 (9.63%) 29 (14.0%) disorders Endocrine disorders 103 (6.92%) 126 (10.0%) 23 (12.2%) 113 (6.54%) 103 (10.3%) 28 (13.5%) Metabolic disorders 147 (9.88%) 492 (39.2%) 62 (32.8%) 173 (10.0%) 488 (48.9%) 82 (39.6%) Malnutrition 3 (0.20%) 5 (0.40%) 2 (1.06%) 4 (0.23%) 5 (0.50%) 4 (1.93%) Genitourinary disorders 46 (3.09%) 113 (9.00%) 21 (11.1%) 48 (2.78%) 103 (10.3%) 22 (10.6%) Hematology disorders 58 (3.90%) 78 (6.21%) 18 (9.52%) 64 (3.70%) 67 (6.72%) 23 (11.1%) Malignant neoplasm 46 (3.09%) 57 (4.54%) 11 (5.82%) 44 (2.54%) 56 (5.62%) 10 (4.83%) Solid organ transplantation 4 (0.27%) 3 (0.24%) 44 (23.3%) 8 (0.46%) 0 (0.00%) 49 (23.7%) Bone marrow transplant 3 (0.20%) 0 (0.00%) 0 (0.00%) 3 (0.17%) 0 (0.00%) 0 (0.00%) Human Immunodeficiency Virus 7 (0.47%) 5 (0.40%) 2 (1.06%) 12 (0.69%) 6 (0.60%) 1 (0.48%) Immunological disorders 18 (1.21%) 26 (2.07%) 14 (7.41%) 33 (1.91%) 21 (2.11%) 12 (5.80%) Table 1: Comorbidity phenotypes identified in the training and test subcohorts. included in the LCA model and patients available for the 2). Patients with the renal-morbidity phenotype showed primary outcome analysis can be found on the online worse oxygenation than patients in the other clusters supplement (eFigure 3). (Figure 1B and eTable 2) but were less prone to having IMV at the time of ICU admission than patients with the low-morbidity phenotype (171 (42%) vs. 1733 Baseline characteristics of the cohort (51.2%), adjusted OR (95% CI) of 0.58 (0.47 to 0.72)), The main baseline characteristics of the 5866 patients while patients in the high-morbidity phenotype had the in the cohort are summarized in Table 2. Briefly, the highest rates of IMV at ICU admission (1166 (56.2%)). median [p25; p75] age was 63 [54; 71] years, 29.6% were Finally, each of the phenotypes showed a characteristic women, and 63.4% were never smokers. The median profile in the blood test (Figure 1C and eTable 2). number of comorbidities was 2 [1; 3], with the most fre- quent comorbidities being hypertension (50.4%), obe- sity (35.5%) and diabetes mellitus (24.9%). In-hospital complications and mortality Specific complication profiles were observed for each phenotype after adjustment for confounding factors Severity and key parameters at ICU admission (Figure 2A and eTable 3). Renal-morbidity and high- The majority of patients presented acute respiratory dis- morbidity phenotypes had a greater number of in-hospi- tress syndrome (ARDS) at ICU admission (95.9%), tal complications than the low-morbidity phenotype. with a median PaO2 to FiO2 ratio of 113 [82; 163]. Both phenotypes showed a greater incidence of myocar- Despite all phenotypes having similar elapsed times dial infarction and ischemia, heart failure, acute kidney from symptom onset to ICU admission (median of 9 [7; injury and anemia. The renal-morbidity phenotype 12] days), renal-morbidity and high-morbidity patients showed a greater incidence of other cardiovascular com- had higher APACHE and SOFA scores than patients plications, such as cardiac arrest, noninfectious shock with the low-morbidity phenotype (Figure 1B and eTable and bleeding. The high-morbidity phenotype had a 6 www.thelancet.com Vol 18 Month July, 2022 Articles Figure 1. Identification of comorbidity phenotypes in the whole CIBERESUCICOVID cohort using Latent Class Analysis. A) Prevalence of comorbidities according to phenotypes. B) Disease severity parameters at ICU admission according to phenotypes. C) Laboratory parameters according to phenotypes. The values in B and C have been standardized. greater incidence of bacterial pneumonia, other infec- Mid-term sequelae: structural and functional lung tions and septic shock. Finally, the low-morbidity phe- impairment notype showed a greater incidence of disseminated Hospital readmissions, complications and sequelae intravascular coagulation (DIC) and pulmonary embo- (including lung structural and functional impairment), lism. adjusted by factors associated with lung damage prior to Regarding in-hospital mortality, both renal-morbid- infection (age, sex, chronic lung disease and asthma), ity and high-morbidity phenotypes had higher in-hospi- are shown in eTable 5. Half of the patients reported per- tal mortality risk than the low-morbidity phenotype, sistent limiting fatigue at 3 and 6 months after dis- with adjusted HRs (95% CIs) of 1.57 (1.34 to 1.84) and charge. All phenotypes showed a high presence of 1.16 (1.05 to 1.28), respectively (Figure 2B). This was pulmonary functional and morphological sequelae at 3 confirmed in a competing risks analysis (eFigure 4). months that persisted over time. Overall, patients with the renal-morbidity phenotype showed worse values than the other phenotypes at the 3 and 6-month follow- Effect of tocilizumab and corticosteroids on mortality up (eTable 5). Tocilizumab was not associated with a reduction in the risk of mortality in any phenotype (eTable 4). Con- versely, corticosteroids showed an overall significant Multimorbidity subphenotypes and its impact on reduction in mortality risk, with this reduction being mortality stronger in the high-morbidity and renal-morbidity phe- Given that all patients with the high-morbidity pheno- notypes although the later not reaching statistical signif- type had two or more comorbidities and 77.5% had icance in a test for interaction (eTable 4). three or more comorbidities, a study of multimorbidity www.thelancet.com Vol 18 Month July, 2022 7 Articles ALL Low-morbidity High-morbidity Renal-morbidity p value n n = 5866 n = 3385 n = 2074 n = 407 Median [p25;p75] Median [p25;p75] Median [p25;p75] Median [p25;p75] or n(%) or n(%) or n(%) or n(%) Sociodemographic data Sex, woman 1732 (29.6%) 1043 (30.8%) 569 (27.5%) 120 (29.5%) 0.031 5859 Age, years 63.0 [54.0;71.0] 60.0 [50.0;68.0] 67.0 [61.0;73.0] 68.0 [60.5;74.0] <0.001 5866 Smoking history <0.001 5386 Non smoker 3416 (63.4%) 2171 (70.3%) 1036 (54.2%) 209 (54.4%) Current 324 (6.02%) 171 (5.53%) 121 (6.33%) 32 (8.33%) Former 1646 (30.6%) 748 (24.2%) 755 (39.5%) 143 (37.2%) Alcohol consumption <0.001 5320 Non consumer 5027 (94.5%) 2930 (95.8%) 1760 (93.3%) 337 (89.9%) Current 194 (3.65%) 94 (3.07%) 83 (4.40%) 17 (4.53%) Former 99 (1.86%) 35 (1.14%) 43 (2.28%) 21 (5.60%) Drug use 0.141 5420 Non consumer 5370 (99.1%) 3067 (98.8%) 1924 (99.5%) 379 (99.2%) Current 25 (0.46%) 18 (0.58%) 6 (0.31%) 1 (0.26%) Former 25 (0.46%) 19 (0.61%) 4 (0.21%) 2 (0.52%) Comorbidities Obesity 2085 (35.5%) 1012 (29.9%) 930 (44.8%) 143 (35.1%) <0.001 5866 Hypertension 2955 (50.4%) 765 (22.6%) 1817 (87.6%) 373 (91.6%) <0.001 5866 Diabetes mellitus (Type I / II) 1458 (24.9%) 162 (4.79%) 1083 (52.2%) 213 (52.3%) <0.001 5866 Chronic heart disease 753 (12.8%) 69 (2.04%) 543 (26.2%) 141 (34.6%) <0.001 5866 Chronic renal disease 423 (7.21%) 29 (0.86%) 0 (0.00%) 394 (96.8%) <0.001 5866 Chronic moderate liver disease 58 (0.99%) 19 (0.56%) 26 (1.25%) 13 (3.19%) <0.001 5866 Chronic mild liver disease 143 (2.44%) 49 (1.45%) 73 (3.52%) 21 (5.16%) <0.001 5866 Chronic neurological disease 321 (5.47%) 110 (3.25%) 177 (8.53%) 34 (8.35%) <0.001 5866 Chronic pulmonar disease 582 (9.92%) 123 (3.63%) 372 (17.9%) 87 (21.4%) <0.001 5866 Asthma 361 (6.15%) 219 (6.47%) 123 (5.93%) 19 (4.67%) 0.314 5866 Dementia 40 (0.68%) 15 (0.44%) 20 (0.96%) 5 (1.23%) 0.019 5866 Rheumatic disease 265 (4.52%) 107 (3.16%) 129 (6.22%) 29 (7.13%) <0.001 5866 Gastrointestinal/ 424 (7.23%) 153 (4.52%) 215 (10.4%) 56 (13.8%) <0.001 5866 pancreatic disorders Endocrine disorders 496 (8.46%) 212 (6.26%) 232 (11.2%) 52 (12.8%) <0.001 5866 Metabolic disorders 1444 (24.6%) 312 (9.22%) 981 (47.3%) 151 (37.1%) <0.001 5866 Malnutrition 23 (0.39%) 8 (0.24%) 8 (0.39%) 7 (1.72%) 0.001 5866 Genitourinary disorders 353 (6.02%) 95 (2.81%) 213 (10.3%) 45 (11.1%) <0.001 5866 Hematology disorders 308 (5.25%) 122 (3.60%) 145 (6.99%) 41 (10.1%) <0.001 5866 Malignant neoplasm 224 (3.82%) 87 (2.57%) 114 (5.50%) 23 (5.65%) <0.001 5866 Solid organ transplantation 108 (1.84%) 13 (0.38%) 0 (0.00%) 95 (23.3%) <0.001 5866 Bone marrow transplant 6 (0.10%) 6 (0.18%) 0 (0.00%) 0 (0.00%) 0.165 5866 Human Immunodeficiency Virus 33 (0.56%) 19 (0.56%) 11 (0.53%) 3 (0.74%) 0.819 5866 Immunological disorders 124 (2.11%) 54 (1.60%) 43 (2.07%) 27 (6.63%) <0.001 5866 Table 2: Characteristics of patients at hospital admission according to phenotypes in the whole CIBERESUCICOVID cohort. Continuous and categorical variables were compared between groups using Kruskall-Wallis test and Chi-squared test, respectively. patterns was performed using a second step of LCA. Six hypertension and others (Figure 3A and eTable 6). Sig- multimorbidity subphenotypes were identified: (i) nificant differences on in-hospital mortality risks were hypertension and chronic lung disease; (ii) hyperten- found when taking subphenotype IV (hypertension and sion and diabetes; (iii) hypertension and chronic heart other metabolic disorders) as a reference, as shown in disease; (iv) hypertension and other metabolic disorders; Figure 3B. Similar results were obtained when using (v) diabetes and other metabolic disorders; and (vi) competing risks model (eTable 7). 8 www.thelancet.com Vol 18 Month July, 2022 Articles Figure 2. Hospital prognosis according to morbidity phenotypes in the whole CIBERESUCICOVID cohort. A) Comparison of the risk of having in-hospital complications between phenotypes. B) In-hospital mortality according to morbidity phenotypes using Cox model. Cox regression model with phenotypes as predictor, and age and sex as confounding factors. Low-morbidity phenotype was used as reference group. Cox model showed a c-statistic of 0.65. 43 patients were excluded from this analysis because of mismatches in the dates of ICU admission and hospital discharge. A total of 808, 784 and 200 patients died during hospitalization in the low-morbidity, high-morbidity and renal-morbidity phenotypes, respectively. Impact of individual comorbidities in a low-morbidity The current study has some key strengths that population enhance its value compared with previous literature. The population of patients in the low-morbidity pheno- First, we performed an in-depth analysis of the indepen- type was best suited for the study of the individual dent impact of comorbidity patterns prior to SARS-CoV- impact of each comorbidity on the risk of intubation 2 infection on the prognosis of critically ill COVID-19 and mortality, which is shown in Figure 4. On the one patients, considering in-hospital outcomes and sequelae hand, CKD, malignant neoplasm and diabetes were sig- and thus being useful for the planning of therapeutic nificantly associated with in-hospital mortality. On the strategies and post-discharge controls. Second, the study other hand, obesity was associated with a higher risk of is based on a large multicentric cohort of critically ill intubation (but not mortality). COVID-19 patients. Third, it includes a broad set of parameters and variables at different time points of the COVID-19 course, including data before ICU admis- Discussion sion, during ICU admission, and up to 6 months after The analysis of the baseline comorbidity patterns of a discharge. This, for instance, allowed us to provide large cohort of critically ill COVID-19 patients allowed some insight into the effectiveness of treatment or the the identification of well-characterized phenotypes with impact of phenotypes on sequelae. Finally, all data were an impact on prognosis, response to treatment mid- thoroughly revised and validated in contrast to registry- term sequelae. Three statistically and clinically signifi- based studies. Despite these strengths, some limitations cant phenotypes were identified: (i) low-morbidity, must be acknowledged. First, the study was limited by including younger patients with few comorbidities; (ii) its observational design. Second, although phenotypes high-morbidity, including patients with two or more were internally validated using training and test subco- comorbidities; and (iii) renal-morbidity, including horts, no formal external validation in a completely patients with CKD and additional comorbidities. As independent cohort was available. Third, the conditional expected, patients with the low-morbidity phenotype independence assumption for LCA did not hold after reported lower ICU mortality than high-morbidity and assessing local dependence based on standardized bivar- renal-morbidity patients. Moreover, an in-depth analysis iate residuals. However, other indicators such as the of the role of specific comorbidities in each of the phe- high average posterior class probabilities, the clinical notypes highlighted the key impact of CKD and, to a interpretability of the latent classes and their reproduc- lesser extent, chronic lung disease in the course of ibility allowed us to believe in the validity of the identi- severe COVID-19. Figure 5 shows a visual overview of fied latent classes. Fourth, data on the administration the study and its main results. timing and dose of pharmacological treatments was not www.thelancet.com Vol 18 Month July, 2022 9 Articles Figure 3. Multimorbidity patterns in patients with high comorbid burden. A) Prevalence of comorbidities according to subpheno- types. B) Impact of subphenotypes on in-hospital mortality. Cox regression model with subphenotypes as predictor, and age and sex as confounding factors. Significance levels were indicate as * if p value<0.05, ** if p value<0.01 and *** if p value<0.001. Cox model showed a c-statistic of 0.63. 10 www.thelancet.com Vol 18 Month July, 2022 Articles Figure 4. Impact of individual comorbidities of patients with low comorbid burden on in-hospital mortality and invasive mechanical ventilation (IMV). Logistic regression models were used to assess the association between comorbidities and IMV risk. Cox propor- tional hazards models were used to assess mortality risk. All models were adjusted for age and sex. The n (%) of subjects having each comorbidity is reported. available. Fifth, a potential impact of therapeutic effort number of patients included in the analyses of sequelae limitation cannot be ruled out, especially in the most was necessarily limited, and the available data were difficult moments of the pandemic. Moreover, therapeu- scarce. Nevertheless, the authors considered it key to tic strategies, which changed through the course of the provide as much data as available, especially linking the pandemic, could have affected the measures of some of acute phase and mid-term follow-up, as this kind of the reported biologic and clinical parameters. Sixth, the study is scarce in the literature. date of complications during hospital stay was not To date, studies aiming to derive phenotypes of hos- recorded, this precluded the use of mortality as a com- pitalized COVID-19 patients have been based on previ- peting risk for complications. Seventh, the sub- ous comorbidities,6−8,15 COVID-19 signs and phenotypes’ cluster analysis among high-morbidity symptoms,8,9,11,12,16,17 variables from blood tests,8,11,13−17 patients could not be internally validated due to con- and chest radiologic features8,10 in an attempt to iden- straints in the number of available subjects. Eighth, the tify groups of patients with distinct risks of ICU admis- number of patients with baseline CKD was not big sion, need for mechanical ventilation, and risk of enough to allow for subgroup analyses, especially con- experiencing clinical complications or death. Among sidering CKD patients that had undergone a solid organ those focused on critically ill patients,11−17 most studies transplantation. Ninth, potentially relevant variables identified two or three phenotypes corresponding to dif- such as ethnicity were not recorded and could not be ferent degrees of severity with a role of intense inflam- potentially used in adjusted models. Finally, the mation when defining the most severe phenotype. The www.thelancet.com Vol 18 Month July, 2022 11 Articles Figure 5. Graphical abstract. Identification of phenotypes and impact on prognosis. Phenotypes identified in the whole CIBERESUCI- COVID cohort by means of Latent Class Analysis based on 17 potentially relevant comorbidities and validated internally using train- ing and test sub-cohorts. The width of the flow lines in each of the phenotypes is proportional to the number of subjects in each time point (all lines are proportional within each phenotype). The width of the flow lines is not proportional between phenotypes (for instance, the renal-morbidity phenotype flow has been over-represented in the sake of a better data visualization). Key charac- teristics of each phenotype: Low-morbidity (n=3385; 58%), younger patients with few comorbidities; High-morbidity (n=2074; 35%), patients with high comorbid burden; Renal-morbidity (n=407; 7%), patients with chronic kidney disease, high comorbidity burden and the worst oxygenation profile. present study differs from the aforementioned studies barrier dysfunction, difficulties associated with main- in being based exclusively on the baseline comorbidity taining the acid-base balance, systemic inflammation patterns of critically ill patients and assessing how dif- and immunodeficiency.24 In contrast, acute renal fail- ferent comorbidity backgrounds influence clinical out- ure, caused directly by SARS-CoV-2 in the renal paren- come, response to available treatment, risk of different chyma or secondary to hemodynamic instability, complications, mortality, and sequelae. inflammatory cytokines or the consequences of ICU Previous research focusing on the comorbidity bur- therapies,25 could be seen as a marker of global vascular den of hospitalized COVID-19 patients has identified damage caused by the infection and thus imply inherent hypertension, diabetes, obesity, COPD, active cancer prognostic value throughout the acute phase. and CKD as having an impact on prognosis and clinical Regarding other comorbidities, on the one hand, outcome.3−5 The present cluster analysis and subse- analyses in the low-comorbidity phenotype showed that quent deepened analyses of the role of comorbidities in only diabetes and active malignancies were associated each phenotype highlighted the key role of CKD. Sub- with higher mortality rates. On the other hand, the anal- jects with baseline CKD, regardless of comorbidity bur- ysis of multimorbidity patterns among patients with the den, experienced the worst in-hospital outcome in high-morbidity phenotype identified subphenotypes terms of complications, either infectious or related to with the highest burden of metabolic diseases as having myocardial infarction and ischemia, and total in-hospi- the best prognosis as opposed to those with pulmonary tal mortality. Even in the low-morbidity phenotype, comorbidities or a wide spectrum of chronic diseases patients with CKD experienced significantly higher in- added to hypertension. Prior studies reported that hospital mortality than those with other or no comorbid- patients with chronic lung disease, especially COPD ities. The impact of baseline CKD on the prognosis of patients, appear to have a predisposition to suffer from COVID-19 could be explained by the effects of kidney severe forms of disease.26−28 This has been related to a failure on general immunity, including intestinal higher expression of ACE-2 receptors in the bronchial 12 www.thelancet.com Vol 18 Month July, 2022 Articles epithelium,29 especially in smokers,30,31 and with the that more than half of COVID-19 survivors experienced impairment of the immune response.32 However, these PASC six months after recovery. Our study, focused on mechanisms seem insufficient to fully explain this ICU patients, shows several relevant findings. First, in result. Perhaps, our findings reporting the poor progno- line with Groff et al., half of the patients, regardless of sis of chronic lung disease in the context of a high their comorbidity phenotype, experience persistent fatigue comorbidity burden highlight the need for the concur- at 3 and 6 months after discharge. Second, patients with rence of other factors not related to lung impairment the high-morbidity and renal-morbidity phenotypes that act as key contributors to adverse outcomes in these showed worse lung function than those with the low-mor- subjects, either by increasing cell infection or by wors- bidity phenotype, with renal-morbidity patients showing ening immune response impairment. the worst results. However, the overall presence of A comparison of key biological parameters between impaired pulmonary function at 3 months after discharge phenotypes showed no significant differences regarding was high in all phenotypes and only the renal-morbidity the severity of respiratory insufficiency at ICU admission, phenotype showed substantial differences in terms of dif- based on PaO2/FIO2, and the degree of inflammation, fusion impairment at 3 and 6 months after discharge, based on C reactive protein. In contrast, significantly which could suggest a differential activation of harmful higher levels of markers of coagulation activation, such as mechanisms in patients with the renal-morbidity pheno- D dimer, were found in the high-morbidity and renal-mor- type (at least at the lung level). Finally, the risk of bidity phenotypes. This could imply a higher degree of experiencing infectious complications, emergency visits or microthrombosis, which has been related to an increased hospital readmissions was significantly higher in renal- risk of in-hospital mortality.33,34 morbidity patients than in those with the low-morbidity As expected, patients with the high-morbidity and phenotype, with patients with the high-morbidity pheno- renal-morbidity phenotypes experienced the highest type falling in between the two. burden of infectious and thrombotic complications. In To conclude, this study identified three well-defined this sense, these patients experienced more cardiovascu- phenotypes in critically ill COVID-19 patients, based on lar events, either ischemic heart disease or myocardial previous comorbidities and with a key role of CKD, and infarction, most probably due to prior subclinical coro- related them to in-hospital complications and mortality, nary plaques being aggravated by coagulation disorder response to standard treatment, and the prevalence of and endothelial damage induced by the virus. In con- sequelae and their progressive remission up to 6 months trast, patients with the low-morbidity phenotype showed after discharge. This highlights the importance of comor- higher pulmonary embolism and disseminated intra- bidity patterns and especially CKD in COVID-19, and vascular coagulation. Despite uncertainties regarding could help clinicians in day-to-day decision making includ- the cause of pulmonary embolism, either local thrombo- ing the management of post-discharge sequelae. sis complications in the gateway of the virus infection or deep vein thrombosis,35 it is likely that its higher inci- dence is observed by the absence of other competing Contributors risks, such as cardiovascular events, leading to the need Conceptualization (IDB, JdB, GT, FB), data curation for ICU admission. Disseminated intravascular coagula- (IDB, CG-P, AM-M), formal analysis (IDB), funding tion is likely caused by viral sepsis, especially in subjects acquisition (AT, FB), investigation (all), methodology struggling to clear the virus, and would not be related to (IDB, JdB, GT, AT, FB), project administration (AT, the comorbidity burden. FB), supervision (AT, FB), writing − original draft Corticosteroids, especially dexamethasone36 and toci- (IDB, JdB, GT), and writing − review & editing (all). lizumab,37 have both been widely used for the treatment Ivan D. Benıtez and Jordi de Batlle have directly of COVID-19. Focusing on corticosteroids, the current accessed and verified the underlying data. Ferran Barbe study showed an overall positive effect on mortality, was responsible for the decision to submit the manu- which was stronger in the high-morbidity and renal- script. morbidity phenotypes (although only the former reached statistical significance in a test for interaction). This contrasts with results by Sinha et al. showing a benefit of corticosteroids only in hyperinflammatory Declaration of interests phenotype patients.16 When looking at tocilizumab, no None declared. effect on mortality was found in any of the phenotypes. This contrasts with previously reported results showing effectiveness if used in the first days of ICU Data sharing statement admission.38,39 Unfortunately, information on treat- An anonymized, de-identified version of the dataset can ment timing was not available in the current study. be made available upon reasonable request to allow In a wide systematic review on persistent post-acute results to be reproduced. A two-year embargo time since sequelae of COVID-19 (PASC), Groff et al.40 reported time of publication will be in place. www.thelancet.com Vol 18 Month July, 2022 13 Articles Acknowledgements Bou, Rafaela Nogueras Salinas, Marta Oc on, Ana The authors are indebted to Maricel Arbones, Maria Ortega, Sergio Ossa, Pablo Pagliarani, Francisco Par- Arguimbau, Raquel Campo, Natalia Jarillo, Javier rilla, Leire Perez Bastida, Purificaci on Perez, Felipe Mu~ noz, Silvia Ortega and Manuel Sanchez for their Perez-Garcıa, Gloria Perez Planelles, Eva Perez Rubio, extensive support in project management and article David Pesta~ na Laguna, Javier Prados, Andres Pujol, preparation. Nuria Ramon Coll, Gloria Renedo Sanchez-Giron, Fer- ran Roche-Campo, Laura Rodriguez, Felipe Rodrıguez de Castro, Silvia Rodrıguez, Covadonga Rodrıguez CIBERESUCICOVID collaborators Ruiz, Jorge Rubio, Alberto Rubio L  opez, Angela Leonor Elena Abril Palomares, Berta Adell-Serrano, Marıa Ruiz-Garcıa, Miriam Ruiz Miralles, Pablo Ryan Mur ua, Aguilar Cabello, Victoria Alcaraz-Serrano, Cesar Alde- Eva Saborido Paz, Ana Salazar Degracia, Inmaculada  coa, Cynthia Alegre, Angela Algaba Calder on, Sergio Salvador-Adell, Miguel Sanchez, Ana Sanchez, Susana  Alvarez,  Antonio Alvarez Ruiz, Ruth Andrea, Maria de Sancho Chinesta, Bitor Santacoloma, Miguel Sanchez, Alba Aparicio, Marta Arrieta, J Ignacio Ayestaran, Joan Maria Teresa Sari~ nena, Esther Sauras-Col on, Marta Ramon Badia, Mariona Badıa, Orville Baez Pravia, Ana Segura Pensado, Lidia Serra, Mireia Serra-Fortuny, Ain- Balan Mari~ no, Bego~na Balsera, Laura Barbena, Enric hoa Serrano Lazaro, Lluıs Servia, Laura Soliva, Carla Barbeta, Tommaso Bardi, Patricia Barral Segade, Marta Speziale, Adrian Tormos, Mateu Torres, Celia Tranque-  Barroso, Jose Angel Berezo Garcıa, Belen Betere, Judit Liberal, Sandra Trefler, Javier Trujillano, Luis Urrelo- Bigas, Rafael Blancas, Marıa Luisa Blasco Cortes, Marıa Cerr on, Estela Val, Luis Valdivia Ruiz, Montserrat Bodi Saera, Marıa Teresa Bouza Vieiro, Leticia Bueno, Vallverdu, Maria Van der Hofstadt Martin-Montalvo, Juan Bustamante-Munguira, Cecilia del Busto Sabela Vara Adrio, Nil Vazquez, Javier Vengoechea, Martınez, David Campi Hermoso, Sandra Campos Clara Vila-Vilardel, Judit Vilanova, Tatiana Villada War- Fernandez, Iosune Cano, Joan Canseco, Pablo Cardina rington, Hua Yang, Minlan Yang, Ana Zapatero. Fernandez, Laura Carrion Garcıa, Sulamita Carvalho, Manuel Castella, Andrea Castellvı, Pedro Castro, Marıa Jose Centelles-Serrano, Ramon Cicuendez Avila,  Catia Cilloniz, Luisa Clar, Cristina Climent, Jordi Codina, Funding Pamela Conde, Sofıa Contreras, Raul de Frutos Parra, Financial support was provided by Instituto de Salud Raul de Pablo Sanchez, Diego De Mendoza, Yolanda Carlos III (CIBERESUCICOVID, COV20/00110), co- Dıaz, Marıa Digna Rivas Vilas, Cristina D olera Moreno, funded by Fondo Europeo de Desarrollo Regional Irene Dot, Pedro Enrıquez Giraudo, Ines Esmorıs (FEDER), “Una manera de hacer Europa”, Centro de Arijon, Teresa Farre Monjo, Javier Fernandez, Carlos Investigacion Biomedica en Red − Enfermedades Res- Ferrando, Albert Figueras, Eva Forcadell-Ferreres, Lor- piratorias (CIBERES) and Donation Program “estar pre-  ena Forcelledo Espina, Enric Franquesa, Angels Furro, parados”, UNESPA, Madrid, Spain. JdB acknowledges Albert Gabarrus, Cristobal Galban, Felipe Garcıa, Bea- receiving financial support from Instituto de Salud Car- triz Garcıa, Emilio Garcıa Prieto, Carlos Garcıa los III (ISCIII; Miguel Servet 2019: CP19/00108), co- Redruello, Amaia Garcıa Sagastume, Maria Luisa funded by the European Social Fund (ESF), “Investing Gascon Castillo, Gemma Goma, Vanesa G omez Casal, in your future”. DdGC acknowledges receiving financial Silvia Gomez, Carmen Gomez Gonzalez, Federico support from Instituto de Salud Carlos III (ISCIII; Gordo, Maria Pilar Gracia, Marıa Jose Gutierrez Fernan- Miguel Servet 2019: CP20/00041), co-funded by the dez, Alba Herraiz, Ruben Herran-Monge, Mercedes European Social Fund (ESF), “Investing in your future”. Ibarz, Silvia Iglesias, Maria Teresa Janer, Gabriel AC acknowledges receiving financial support from Jimenez, Mar Juan Dıaz, Karsa Kiarostami, Juan I Lazo Instituto de Salud Carlos III (ISCIII; Sara Borrell 2021:  Alvarez, Miguel Leon, Alexandre Lopez-Gavın, Desiree CD21/00087). None of the funding sources had a role Macias Guerrero, Nuria Mamolar Herrera, Rafael in the study design; in the collection, analysis, and inter- Ma~ nez Mendiluce, Cecilia L Mantellini, Gregorio Marco pretation of data; in the writing of the report; and in the Naya, Iris Marco Barcos, Pilar Marcos, Enrique Marmol decision to submit the article for publication. Peis, Marta Martın Cuadrado, Paula Martın Vicente, Marıa Martınez, Carmen Eulalia Martınez Fernandez, Maria Dolores Martınez Juan, Basilisa Martınez Pala- cios, Juan Fernando Masa Jimenez, Joan Ramon Transparency statement Masclans, Emilio Maseda, Eva Marıa Menor Fernandez, The leading authors affirm that the manuscript is an Priscila Metora Banderas, Mar Miralbes, Josman Mon- honest, accurate, and transparent account of the study clou, Juan Carlos Montejo-Gonzalez, Neus Montserrat, being reported; that no important aspects of the study Marıa Mora Aznar, Dulce Morales, Sara Guadalupe have been omitted; and that any discrepancies from the Moreno Cano, David Mosquera Rodrıguez, Rosana study as originally planned (and, if relevant, registered) Mu~ noz-Berm udez, Jose Marıa Nicolas, Ramon Nogue have been explained. 14 www.thelancet.com Vol 18 Month July, 2022 Articles Data availability statement 16 Sinha P, Furfaro D, Cummings MJ, et al. Latent class analysis Excerpts of relevant data can be available from the corre- reveals COVID-19-related ARDS subgroups with differential responses to corticosteroids. Am J Respir Crit Care Med. 2021:1–65. sponding author on reasonable request. However, a two 17 Vasquez CR, Gupta S, Miano TA, et al. Identification of distinct years embargo is foreseen to allow authors to fully com- clinical subphenotypes in critically ill patients with COVID-19. Chest. 2021;160. https://doi.org/10.1016/j.chest.2021.04.062. plete their analysis and publication plans. 18 Torres A, Arguimbau M, Bermejo-Martın J, et al. CIBERESUCICO- VID: a strategic project for a better understanding and clinical management of COVID-19 in critical patients. Arch. Bronconeumol. 2021;57. https://doi.org/10.1016/j.arbres.2020.10.021. Supplementary materials 19 Inker LA, Eneanya ND, Coresh J, et al. New creatinine- and cystatin C−based equations to estimate GFR without race. N Engl J Med. Supplementary material associated with this article can 2021;385. https://doi.org/10.1056/nejmoa2102953. be found in the online version at doi:10.1016/j. 20 Bandeen-roche K, Miglioretti DL, Zeger SL, Rathouz PJ. Latent var- iable regression for multiple discrete outcomes. J Am Stat Assoc. lanepe.2022.100422. 1997;92. https://doi.org/10.1080/01621459.1997.10473658. 21 NAGIN D. Group-based modeling of development. 2005. Chapter 5. DOI:10.4159/9780674041318. References 22 Austin PC, Lee DS, Fine JP. Introduction to the analysis of survival 1 Huang C, Wang Y, Li X, et al. Clinical features of patients infected data in the presence of competing risks. Circulation. 2016;133. with 2019 novel coronavirus in Wuhan, China. Lancet. 2020;395. https://doi.org/10.1161/CIRCULATIONAHA.115.017719. https://doi.org/10.1016/S0140-6736(20)30183-5. 23 Noordzij M, Leffondre K, Van Stralen KJ, Zoccali C, Dekker FW, 2 Zhang JJY, Lee KS, Ang LW, Leo YS, Young BE. Risk factors for Jager KJ. When do we need competing risks methods for survival severe disease and efficacy of treatment in patients infected with analysis in nephrology? Nephrol Dial Transplant. 2013;28. https:// COVID-19: a systematic review, meta-analysis, and meta-regression doi.org/10.1093/ndt/gft355. analysis. Clin Infect Dis. 2020;71:2199–2206. 24 Kurts C, Panzer U, Anders HJ, Rees AJ. The immune system and 3 Liang W, Liang H, Ou L, et al. Development and validation of a kidney disease: basic concepts and clinical implications. Nat Rev clinical risk score to predict the occurrence of critical illness in hos- Immunol. 2013;13:738–753. pitalized patients with COVID-19. JAMA Intern Med. 25 Gabarre P, Dumas G, Dupont T, Darmon M, Azoulay E, Zafrani L. 2020;180:1081–1089. Acute kidney injury in critically ill patients with COVID-19. Inten- 4 Gupta RK, Harrison EM, Ho A, et al. Development and validation sive Care Med. 2020;46. https://doi.org/10.1007/s00134-020- of the ISARIC 4C deterioration model for adults hospitalised with 06153-9. COVID-19: a prospective cohort study. Lancet Respir Med. 2021;9. 26 Guan W, Liang W, Zhao Y, et al. Comorbidity and its impact on https://doi.org/10.1016/S2213-2600(20)30559-2. 1590 patients with COVID-19 in China: a nationwide analysis. Eur 5 Grasselli G, Zangrillo A, Zanella A, et al. Baseline characteristics Respir J. 2020;55. https://doi.org/10.1183/13993003.00547-2020. and outcomes of 1591 patients infected with SARS-CoV-2 admitted Supp1. to ICUs of the lombardy region, Italy. JAMA - J Am Med Assoc. 27 Alqahtani JS, Oyelade T, Aldhahir AM, et al. Prevalence, severity 2020;323. https://doi.org/10.1001/jama.2020.5394. and mortality associated with COPD and smoking in patients with 6 Wang X, Jehi L, Ji X, Mazzone PJ. Phenotypes and subphenotypes COVID-19: A rapid systematic review and meta-analysis. PLoS of patients With COVID-19: a latent class modeling analysis. Chest. One. 2020;15. https://doi.org/10.1371/journal.pone.0233147. 2021;159:2191–2204. 28 Feng Y, Ling Y, Bai T, et al. COVID-19 with different severities: a 7 Lusczek ER, Ingraham NE, Karam BS, et al. Characterizing multicenter study of clinical features. Am J Respir Crit Care Med. COVID-19 clinical phenotypes and associated comorbidities and 2020;201. https://doi.org/10.1164/rccm.202002-0445OC. complication profiles. PLoS One. 2021;16:1–18. 29 Leung JM, Yang CX, Tam A, et al. ACE-2 expression in the small 8 Gutierrez-Gutierrez B, del Toro MD, Borobia AM, et al. Identifica- airway epithelia of smokers and COPD patients: implications for tion and validation of clinical phenotypes with prognostic implica- COVID-19. Eur. Respir. J. 2020;55. https://doi.org/10.1183/ tions in patients admitted to hospital with COVID-19: a 13993003.00688-2020. multicentre cohort study. Lancet Infect Dis. 2021;21:783–792. 30 Cai G, Bosse Y, Xiao F, Kheradmand F, Amos CI. Tobacco smoking 9 Rubio-Rivas M, Corbella X, Mora-Lujan JM, et al. Predicting clini- increases the lung gene expression of ACE2, the Receptor of SARS- cal outcome with phenotypic clusters in covid-19 pneumonia: an CoV-2. Am. J. Respir. Crit. Care Med. 2020;201. https://doi.org/ analysis of 12,066 hospitalized patients from the spanish registry 10.1164/rccm.202003-0693LE. semi-covid-19. J Clin Med. 2020;9:1–19. 31 Zhang H, Rostami MR, Leopold PL, et al. Expression of the SARS- 10 Robba C, Battaglini D, Ball L, et al. Distinct phenotypes require dis- CoV-2 ACE2 receptor in the human airway epithelium. Am J Respir tinct respiratory management strategies in severe COVID-19. Crit Care Med. 2020;202. https://doi.org/10.1164/rccm.202003- Respir. Physiol. Neurobiol. 2020;279. https://doi.org/10.1016/j. 0541OC. resp.2020.103455. 32 Bhat TA, Panzica L, Kalathil SG, Thanavala Y. Immune dysfunc- 11 Azoulay E, Zafrani L, Mirouse A, Lengline E, Darmon M, Chevret tion in patients with chronic obstructive pulmonary disease. In: S. Clinical phenotypes of critically ill COVID-19 patients. Intensive Annals of the American Thoracic Society. 2015. https://doi.org/ Care Med. 2020;46:1651–1652. 10.1513/AnnalsATS.201503-126AW. 12 Bos LDJ, Sjoding M, Sinha P, et al. Longitudinal respiratory sub- 33 Ye W, Chen G, Li X, et al. Dynamic changes of D-dimer and neu- phenotypes in patients with COVID-19-related acute respiratory trophil-lymphocyte count ratio as prognostic biomarkers in distress syndrome: results from three observational cohorts. Lancet COVID-19. Respir Res. 2020;21. https://doi.org/10.1186/s12931- Respir Med. 2021;2600. https://doi.org/10.1016/s2213-2600(21) 020-01428-7. 00365-9. 34 Qeadan F, Tingey B, Gu LY, Packard AH, Erdei E, Saeed AI. Prog- 13 Chen H, Zhu Z, Su N, et al. Identification and prediction of novel nostic values of serum ferritin and d-dimer trajectory in patients clinical phenotypes for intensive care patients with SARS-CoV-2 with covid-19. Viruses. 2021;13. https://doi.org/10.3390/v13030419. pneumonia: an observational cohort study. Front Med. 2021;8:1–9. 35 Suh YJ, Hong H, Ohana M, et al. Pulmonary embolism and 14 Dupont T, Caillat-Zucman S, Fremeaux-Bacchi V, et al. Identifica- deep vein thrombosis in COVID-19: a systematic review and tion of distinct immunophenotypes in critically ill coronavirus dis- meta-analysis. Radiology. 2021;298. https://doi.org/10.1148/ ease 2019 patients. Chest. 2021;159. https://doi.org/10.1016/j. RADIOL.2020203557. chest.2020.11.049. 36 Horby P, Lim WS, Emberson J, et al. Effect of dexamethasone in 15 Rodrıguez AH, Ruiz-Botella M, Martın-Loeches I, et al. Deploying hospitalized patients with COVID-19 − preliminary report. N Engl unsupervised clustering analysis to derive clinical phenotypes and J Med. 2020. https://doi.org/10.1101/2020.06.22.20137273. risk factors associated with mortality risk in 2022 critically ill 37 Abani O, Abbas A, Abbas F, et al. Tocilizumab in patients admitted patients with COVID-19 in Spain. Crit Care. 2021;25:1–15. to hospital with COVID-19 (RECOVERY): a randomised, www.thelancet.com Vol 18 Month July, 2022 15 Articles controlled, open-label, platform trial. Lancet. 2021;397. https://doi. 39 The REMAP-CAP Investigators*. Interleukin-6 receptor antago- org/10.1016/S0140-6736(21)00676-0. nists in critically ill patients with covid-19. N Engl J Med. 38 Gupta S, Wang W, Hayek SS, et al. Association between early treat- 2021;384:1491–1502. ment with tocilizumab and mortality among critically ill patients 40 Groff D, Sun A, Ssentongo AE, et al. Short-term and long-term rates of with COVID-19. JAMA Intern Med. 2021;181. https://doi.org/ postacute sequelae of SARS-CoV-2 infection. JAMA Netw Open. 10.1001/jamainternmed.2020.6252. 2021;4. https://doi.org/10.1001/jamanetworkopen.2021.28568. 16 www.thelancet.com Vol 18 Month July, 2022