key: cord-0958353-wsxcmmcq
authors: Maibach, M. A.; Allam, A.; Hilty, M. P.; Perez Gonzales, N. A.; Buehler, P. K.; Wendel Garcia, P. D.; Brugger, S. D.; Ganter, C. C.; The CoViD-19 ICU-Research Group Zurich,; The RISC-19-ICU Investigators,; Krauthammer, M.; Schuepbach, R. A.; Bartussek, J.
title: Timing COVID-19 - Synchronization of longitudinal patient data to the underlying disease progression using CRP as a temporal marker
date: 2020-06-12
journal: nan
DOI: 10.1101/2020.06.11.20128041
sha: a87020e2b4a7e4f16805ea322ae784ad0c49e76c
doc_id: 958353
cord_uid: wsxcmmcq

Advances in medical technology and IT infrastructure have led to increased availability of continuous patient data that allows to investigate the longitudinal progression of novel and known diseases in unprecedented detail. However, to accurately describe any underlying pathophysiology with longitudinal data, the individual patient trajectories have to be synchronized based on temporal markers. In this study, we use longitudinal data from 28 critically ill ICU COVID-19 patients to compare the commonly used alignment markers "onset of symptoms", "hospital admission" and "ICU admission" with a novel objective method based on the peak value of inflammatory marker C-reactive protein (CRP). By applying our CRP-based method to align the progression of neutrophils and lymphocytes, we were able to define a pathophysiological window that allowed further mortality risk stratification in our COVID-19 patient cohort. Our data highlights that proper synchronization of patient data to the underlying pathophysiology is crucial to differentiate severity subgroups and to allow reliable interpatient comparisons.

The rapid spread of corona virus disease 19 (COVID-19) imposes a heavy burden on public health systems around the world. A substantial number of patients show a severe disease progression possibly caused by endotheliitis, gas diffusion impairment and organ ischemia [1, 2] . Current research efforts focus on the identification of predictive indicators that allow closer supervision and targeted intervention in high-risk patients. As a hyper-activated immune response might act as a driving factor for severe , ratios between neutrophils and lymphocytes (NLR) [3] , lymphocyte counts alone [4] and elevation of specific cytokines among other laboratory values [3, [5] [6] [7] have been proposed as markers for initial patient risk assessment and stratification. Currently, most studies solely compare measurements taken at hospital or intensive care unit (ICU) admission, neglecting the enormous potential of continuous longitudinal data obtained throughout hospitalization. This is especially detrimental for the most severe patients, as this high mortality risk group could benefit the most from a more detailed separation into different disease progression subgroups. However, the pooling of longitudinal data requires a temporal marker to align individual patient trajectories. In this pilot study, we compare different alignment methods and show that C-reactive protein (CRP) can be used to synchronize the individual patient trajectories with the underlying pathophysiology.

To date, only few studies show longitudinal data of COVID-19 patients, and single time point patient comparisons are often made based on clinical parameters such as onset of symptoms, hospital or ICU admission [3] [4] [5] [6] [7] . Alignment based on these values could potentially introduce unnecessary interpatient variation due to human bias and circumstantial parameters such as hospital capacity, resource availability or accessibility, and could eventually result in a clouding of the true underlying disease progression. Comparison of individual patient trajectories in our cohort of 28 critically ill COVID-19 patients admitted to the ICU of the University Hospital Zurich (Suppl. Table 1 ) already revealed considerable inter-patient variability (Fig. 1 a) : Out of the 28 patients, 8 were directly transferred to the ICU upon hospital admission and 5 additional patients were transferred to the ICU only one day after hospital admission. Based on this data alone, it is evident that interpatient comparison at hospital or ICU admission was biased in our ICU COVID-19 patient cohort. Likewise, onset of symptoms showed a high variation (7.65 ± 8.49 days) and was occasionally missing. Group comparisons based on these clinical markers might result in the description of false differences, for example by comparing patients in late disease stages with patients in early disease stages, or occlusion of actual differences due to temporal misalignment ( Fig. 1 b) . These problems are encountered by most medical centers and researchers alike and highlight the necessity for an objective synchronization marker that aligns individual patient trajectories to the underlying pathophysiology. An ideal disease timer should provide both an early indication of disease progression and should be measured routinely in most hospital settings. Previous patient cohort studies have correlated the serum levels of the inflammatory cytokine interleukin 6 (IL-6), myoglobin and cardiac troponin at hospital admission with COVID-19 severity [6] . However, these values were not measured on a daily basis around ICU admission both in our cohort (38.2%, 59.7%, 62.7% respectively) and in the international RISC-19-ICU registry cohort of critically ill COVID-19 patients (14.6%, 9.6%, 30.0% in Switzerland and 15.3%, 6.7%, 28.2% internationally), thereby making them poor candidates for longitudinal alignment (Fig. 1 c) . Instead, the acute phase inflammatory marker CRP was measured routinely around ICU admission both in our cohort (98.2%) and in the RISC-19-ICU registry cohort (86.9% in Switzerland, 74.8% internationally). CRP is under direct transcriptional control of IL-6, but shows a wider peak, making it more likely to be recorded by daily measurements and when the patient is hospitalized [8] . In contrast to other frequently measured laboratory values such as hematological cell counts or creatinine, we found that most patients had a distinct CRP maximum around ICU admission in our cohort, indicating a correlation with COVID-19 severity and progression ( Fig. 1 d) . Some patients showed further CRP maxima during their ICU stay, probably resulting from coinfections or secondary damage [8] . We hypothesize that the first local CRP maximum (CRP max ) is primarily related to COVID-19 progression and can therefore be used to synchronize individual patient trajectories to the underlying pathophysiology. In our patient cohort, alignment based on CRP max decreased both interpatient variability in the longitudinal CRP curve (Fig. 1 d) and the variability of other laboratory values such as total leukocyte and relative neutrophil and lymphocyte counts to a similar extent than the clinically based ICU admission alignment (Suppl. Fig. 1 ). Our primary goal to accurately align longitudinal patient data was to further differentiate between the most severe ICU COVID-19 patients. To test whether CRP max -based synchronization improves patient stratification in our ICU patient cohort, we retrospectively defined three severity subgroups: (1) deceased ICU patients (n=6), (2) discharged ICU patients that had been mechanically ventilated (n=13) and (3) discharged ICU patients that had been spontaneously breathing while in the ICU (n=9). CRP peak values were more than threefold higher in the mechanically ventilated patient subgroups (mean±SD, 346±147 mg/L) as compared to the spontaneously breathing subgroup (99±74 mg/L), but did not differ from the deceased subgroup (338±106 mg/L) (Fig 1 e) . This lack of distinction is reflected in all alignment methods. In accordance to current literature, we further assessed the longitudinal progression of relative neutrophils and lymphocyte counts (Suppl. Fig. 2 ) and the ratio thereof (NLR, Fig. 1 e) in the three severity subgroups [3, 4] . While both admission-based and CRP max -based alignment improved subgroup separation, only CRP max -based synchronization revealed a distinct NLR turning point, occurring simultaneously with CRP max , thereby providing a window for maximal subgroup distinction. A linear mixed effect model [9] employing subgroup and time as fixed effects and per-patient random slopes as random effects confirmed a difference between the subgroups and the measured time points in a window of ±4 days around CRP max (p<0.01, Suppl. Table 1 ), whereas the time wise difference was not detected in the data shifted by ICU admission (Suppl . Table 2 ). Similarly, when comparing the subgroups in single time points of each alignment method, only the CRP max -based synchronization resulted in a significant difference between the two most severe patient subgroups ( Fig. 1 f) . These findings strengthen our CRP maxbased synchronization as a potential tool for patient alignment and subgroup stratification.

. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted June 12, 2020. . https://doi.org/10. 1101 /2020 In a last step, we explored whether different patient synchronization methods might have an impact on future outcome prediction using machine learning techniques (Suppl. Fig. 3 ). We generated two feature vectors for each patient containing the mean values of CRP, relative neutrophils and lymphocytes of a time window anchored on either ICU admission or on CRP max (Suppl. Fig. 3 upper panel). Using a stratified 5-fold cross validation logistic regression model, we found that the CRP max anchoring increased the overall prediction accuracy by 9.6% and F1-macro score by 51.8% (accuracy 0.68 ± 0.22, F-score 0.668 ± 0.23) as compared to the ICU admission anchoring (accuracy 0.62 ± 0.10, F-score 0.44 ± 0.13) (Suppl. Fig. 3 b-c, window 1). Similarly, the corresponding confusion matrices indicated a higher accuracy in distinguishing between the most severe subgroups of mechanically ventilated and deceased ICU patients (Suppl. Fig. 3 d-e). Collectively, this data suggests that pathophysiological synchronization of longitudinal patient data has the potential to improve mortality-risk stratification and subgroup distinction of severe ICU patients both in a clinical setting and for research purposes.

In this pilot study containing 28 critically ill COVID-19 patients, we demonstrated that longitudinal data synchronization based on the inflammatory marker CRP reduces interpatient variability at least to an equal extend as the ICU admission based alignment. Both "onset of symptoms" and "hospital admission" were poor temporal markers, leading to increased variability, occlusion of subgroup differences and, in case of "onset of symptoms", to patient exclusions due to unclear data. The interpretation and translation of noteworthy symptoms from patients to clinicians make "onset of symptoms" a highly subjective value for patient synchronization, which is reflected in our data and early reports of exaggerated incubation periods until onset of disease [5, 10] . While ICU admission is a consistent clinical marker in our monocentric study, this might not be the case when comparing patients from different hospitals with less stringent or deviating ICU admission criteria, resources or ICU capacity. Furthermore, COVID-19 associated symptoms might not be the primary reason for ICU admission in some patients and, obviously, this temporal marker cannot be applied to non-ICU patients. In contrast, our findings suggest that CRP can serve as an objective synchronization marker that allows alignment of disease trajectories independent of hospital specific policies, which is of special value for multicenter studies. In line with previous studies, our subgroup analysis of both CRP max and ICU aligned data reproduced the COVID-19 severity markers: neutrophilia, lymphocytopenia and the ratio thereof [3] [4] [5] [6] [7] . However, only CRP max -based longitudinal alignment improved distinction between the most severe subgroups of mechanically ventilated patients and deceased patients. Although this pilot study relies on a small cohort, our data suggests a central role for CRP in the timing of COVID-19 immunopathology by marking the turning point of longitudinal NLR dynamic and thereby providing a window for maximal subgroup distinction. Interestingly, CRP itself has immune-modulating functions such as complement activation, regulation of apoptosis and cellular processes of both neutrophils and monocyte-derived cells [8] . Although an elevation of CRP is generally associated with bacterial rather than viral infections [8, 11, 12] , elevated CRP levels have been observed in COVID-19 patients as well as in severe progression of other respiratory viral diseases such as influenza [3, 4, 6, 13, 14] . It is tempting to speculate that elevation of CRP in severe respiratory viral infections marks a shift from a more localized inflammation of the lungs to a multi-organ systemic immune response. Digitalization of modern medicine has led to increased availability of continuous patient data that should be used to describe and define longitudinal disease progression and pathophysiology of novel and known diseases alike. Our data highlights that proper synchronization of longitudinal patient trajectories to the underlying pathophysiology is crucial to differentiate severity subgroups and allow reliable interpatient comparisons.

as researchers for their efforts in the battle against SARS-CoV-2. This research was supported from non-restricted grants to Reto A. Schuepbach.

The datasets analyzed and code generated during the current study are available from the corresponding authors on reasonable request. The RISC-19-ICU data set (NCT04357275) is available on reasonable request (https://www.risc-19-icu.net/main-page/how-to-participate). 

The authors declare no competing interests.

. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted June 12, 2020. . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted June 12, 2020. . https://doi.org/10.1101/2020.06.11.20128041 doi: medRxiv preprint . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted June 12, 2020. . https://doi.org/10.1101/2020.06.11.20128041 doi: medRxiv preprint Supplementary Figure 1 Suppl. Fig. 1 Dynamic changes of CRP, leukocytes and relative neutrophils and lymphocytes. Longitudinal CRP, leukocyte, neutrophils and lymphocytes shifted relative to symptom onset (left), hospital admission (middle, left), ICU admission (middle, right) or first local CRP maximum (right). Synchronization based on onset of symptoms resulted in the exclusion of two deceased patients due to unclear data. Data is shown as median ± MAD. Curves are cut-off when data of fewer than three patients was available. The respective patient numbers are shown in the bottom panels.

. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted June 12, 2020. . https://doi.org/10.1101/2020.06.11.20128041 doi: medRxiv preprint Supplementary Figure 2 Suppl. Fig. 2 Dynamic changes of leukocytes, relative and absolute lymphocytes and neutrophils of severity subgroups. Longitudinal data is synchronized based on onset of symptoms (left) hospital admission (middle, left), ICU admission (middle, right) or first local CRP maximum (right). Synchronization based on onset of symptoms resulted in the exclusion of two deceased patients due to unclear data. Data is shown as median ± MAD. Curves are cut-off when data of fewer than three patients was available. The respective patient numbers are shown in the bottom panels.

. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted June 12, 2020 . . https://doi.org/10.1101 /2020 Supplementary Figure 3 Suppl. Fig. 3 Timer-based risk stratification could improve outcome prediction. a Graphical representation of the data set generation and the applied 5-fold cross validation model. Two feature vectors were generated for each patient containing the mean values of CRP, relative neutrophils and lymphocytes of a time window anchored on either ICU admission or on CRP max (upper panel). We followed a stratified 5-fold cross-validation scheme, where each fold was defined as a distinct 80%-20% train-test split. Within each fold, hyper-parameter selection was performed in the training set with a stratified 4-fold cross validation. For each fold multiple logistic regression models were trained using varying hyper-parameters such as regularization type (l ) and with or without class weighting. The best models as determined by F1-macro score on the 4-fold cross validation were then tested on the test split. Since our retrospective patient classification was done at ICU discharge, only values before outcome classification were considered for the construction of the feature vectors. This lead to the exclusion of 2 patients from the spontaneously breathing subgroup. b -c Mean accuracy performance (b) and mean Macro-f1 score (c) of ICU admission or CRP max anchoring. Data is reported as mean ± SD. d -e Confusion matrices constructed from the best performing trained model of each fold using the test data from all five folds of the ICU admission anchored (d) or CRP max anchored (e) window size 1 data set.

. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted June 12, 2020. . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted June 12, 2020. . https://doi.org/10.1101/2020.06.11.20128041 doi: medRxiv preprint . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted June 12, 2020. . https://doi.org/10. 1101 /2020 

Austria: Department for Anesthesiology and Intensive Care

Czech Republic: Klinika anesteziologie perioperacni a intenzivni mediciny

Italy: Anesthesia and Intensive Care Unit

UO Anestesia e Terapia Intensiva

MD); Department of Anesthesia and Intensive Care Medicine

Chiara Cogliati, MD); Division of Anesthesia and Intensive Care

Netherlands: Department of Intensive Care Medicine

Medical Intensive Care Unit, Hospital Clinic de Barcelona

Acute Critical Cardiac Care Unit, Hospital Clinic de Barcelona

Liver Intensive Care Unit, Hospital Clinic de Barcelona

Respiratory Intensive Care Unit, Hospital Clinic de Barcelona

Departement for intensive care medicine

Departement of Anesthesiology and Intensive Care Medicine, Cantonal Hospital St

Frauenfeld (Alexander Dullenkopf, MD; Lina Petersen, MD); Division of Neonatal and Pediatric Intensive Care

Paediatric Intensive Care Unit, Children's Hospital of Eastern Switzerland

Spitalzentrum Oberwallis

Nuria Zellweger); Department of Intensive Care Medicine

Service

United Kingdom: Harefield Hospital

We thank Catharina Giese, Patrick Hirschi, the RDSC and the PDMS group of the University Hospital Zurich for their continued support. We further thank all the public health and essential workers as well