CoLab-score and viability PCR for SARS-CoV-2: routine blood test algorithm proven by elimination of viable SARS-CoV-2 to identify patients in order to facilitate discharge and/or de-isolation

Problem:
Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) disease 2019
(COVID-19) is a pandemic with over 110 million cases and over two million
deaths worldwide.(1) The initial clinical symptoms for COVID-19 are nonspecific
and similar to other seasonal respiratory viral diseases, which encompasses
fever, dyspnea, dry cough and fatigue. Many countries, including the
Netherlands, are struggling to control COVID-19 outbreaks, especially in the
detection of silent infections in the pre- or asymptomatic patient that can
contribute to transmission.(2) Empirical studies have indicated that
individuals may be also infectious during the presymptomatic phase(3).
Reverse Transcription-Polymerase Chain Reaction (RT-PCR)-based methodologies
are the gold standard in confirming that the individual presenting with
COVID-19 symptoms has active viral shedding of SARS-CoV-2.(2) These techniques
are based on the detection of the presence of viral RNA and are widely
implemented to diagnose the disease. Sampling error by erroneous swapping may
lead to false negative test-results. Conversely, RNA can also be detected while
the patient has overcome the virus and is still recovering. This RNA probably
does not reflect viral reproduction but are remnants of viral RNA (so-called
non-viable RNA).(4) This leads to false positive results that impact care as
patients can only be transferred from COVID-ICU or -ward beds to regular wards
if such tests are negative. Next to this, the sensitivity of testing varies
with timing of testing relative to exposure. One modeling study estimated
sensitivity at 33% 4 days after exposure, 62% on the day of symptom onset, and
80% 3 days after symptom onset.(5) Factors contributing to false-negative test
results include the adequacy of the specimen collection technique, time from
exposure, and specimen source. Furthermore, current and future viral changes
could affect viral based diagnostics.(6-8)

PCR testing: non-viable versus viable RNA
According to the current Dutch guidelines (9) COVID-19 patients are placed in
isolation to prevent spread till they test consecutively two times negative in
the SARS-CoV-2 RT-PCR and/or show presence of anti-IgG*s against SARS-CoV-2 in
serological test. Taken into consideration the high prevalence of COVID-19
during the past waves, the Intensive Care Units (ICU) are overwhelmed with
COVID-19 patients, and each patient in isolation has an enormous organizational
and financial impact on the healthcare system. Availability of ICU wards have
appeared the major bottleneck in these situations, not only limiting care for
COVID-19 patients, but particularly leading to downscaling of non-COVID-19,
regular health care, thus limiting care for all hospital patients. Especially
treatment and taking care of the patients in isolation coincides with physical
and psychological load for the health care workers. Next to this, it has also
an enormous impact on the patient and their relatives due to the effects that
belong to the restrictions accompanied with isolation. A better flow of ICU
patients that can leave isolation because they are really COVID-19 negative
will lead to a relief in the burdening in ICU capacity and that of the logistic
challenges between ICU isolation rooms and non-isolation rooms, but also
between ICU*s and normal wards. This must be done in a safe and responsible
manner without contaminating and potentially infecting other patients and
possibly non-vaccinated healthcare workers. Despite the extensive vaccination
strategy that currently takes place in the Netherlands, this problem will
persist in the future, because of the constant presence of vulnerable patient
groups that likely remain susceptible to the virus (e.g., immunocompromised
patients)
Techniques based on highly sensitive viral-RNA detection may thus hamper
hospital care logistics, and this leads to one of the most important issues
during the past pandemic waves. For that reason, we are searching for other
diagnostic factors, such as biomarkers that give information about the
host-response, leading to an objective determination whether a patient has
overcome SARS-CoV-2 virus and can be safely discharged from isolation is a
logical additional step. Validation of this strategy with an innovative method
that detects only viable RNA can be a solution to definitely prove that these
patients are really non-infectious. The application of a so-called viability
PCR has been successfully applied to bacterial infections by project group
members.(4) This method can also be used for viral infections to determine
viral activity.(10) The project group has now adapted the method for
application to SARS-CoV-2 and has shown that 99% of free viral RNA can be
eliminated, allowing detection of only infectious virus (unpublished data).

Blood testing in COVID
From presentation at the Emergency Department till discharge from the hospital,
many blood samples are drawn for routine laboratory diagnostics and monitoring
disease state. Due to the large number of patients admitted to the hospital for
COVID-19 related complaints, a lot of laboratory data from individual patients
are stored in the different hospital systems, leading to the generation of a
large amount of laboratory data. As shown in the living systematic review from
Wynants et al(11), a considerable number of prediction models for COVID-19 has
been published until recently, and laboratory parameters are often an important
part of these prediction models. Several studies have already shown that
COVID-19 is characterized by biochemical as well as hematological changes in
peripheral blood.(12-14) Although the current proposal focusses on the
development of a working algorithm, the extent of these biochemical and
hematological changes can also provide information about the severity and
prognosis of COVID-19 disease course. Recently, we showed that already at an
early stage of COVID-19 hemocytometric changes could be detected in
immunocompetent leukocytes present in peripheral blood and these changes
correlated with the severity of the infection.(14) With regard to diagnostics,
the CoLab score was developed in a multicenter study and validated based on the
use of combination of biochemical and hematological parameters only.(15) With
this study it is shown that the host reaction to SARS-CoV-2 infection could be
measured early in blood.

CoLab-score
The CoLab-score was developed by members of our project group(15), using an
adaptive LASSO-regression technique.(16) It is a mathematical algorithm that
needs the input of 10 numeric values of routine-laboratory parameters in
addition to age of the patient, and results in a linear predictor (LP). The
resulting score has a very high negative predictive value in the developmental
study (NPV; 99.7%; sensitivity 98.4%) and this was almost the same in the
validation cohort (99.5%; sensitivity 96.9%).(15) This algorithm was developed
and validated to exclude COVID-19 in patients presenting at the Emergency
Department. Apparently, the strength of the high NPV derived from this
algorithm is driven by the absence of specific COVID-19 related biochemical and
hematological changes in peripheral blood. In the validation study of the CoLab
algorithm, it became apparent that the score decreased when the time between
the onset of COVID-19 symptoms and ED presentation increased.(15) In a recent
study that we performed in a screening-setting, we demonstrated that the
CoLab-score can also be used to rule out an infection in healthcare workers
that have been tested because of COVID-19 related complaints or patients in a
pre-surgically setting. In this study, we demonstrated that the cut-off to be
used in the CoLab-algorithm in the screening setting was different from that of
patients presenting at the ED (manuscript in preparation). This could be
explained by the fact that in the screening setting the duration of the
infective period is shorter, the complaints are milder and consequently the
host-response also less pronounced. Next to this, we noticed that in several
patients the outcome of the CoLab-algorithm was not negative, while the
SARS-CoV-2 RT-PCR was definitive negative, and a week later when the patient
had a retest, the outcome of the algorithm had worsened and at that time the
RT-PCR test becomes positive. It appears that the outcome of the CoLab-score is
dynamic and follows the host-reaction. For that reason, our hypothesis is that
these biochemical and hematological changes rapidly return to normal values
after the host clears the infection, resolving COVID-19.

In summary: currently, a gold standard to determine when a patient with
COVID-19 is no longer infectious is lacking. The insufficient predictive value
of current standard PCR testing hampers to take the person out of isolation,
leading to prolonged and unnecessary use of (ICU- and non-ICU) isolation
capacity. An earlier cessation of isolation of COVID-19 patients will obviously
benefit the patient due to the social and emotional impact of regained contact
with their relatives and healthcare providers.
Because the algorithm to derive the CoLab-score has a high NPV for ruling out
COVID-19, we want to investigate in the present study whether this algorithm
can be used to determine when a COVID-19 patient has indeed overcome the
infection. This dynamic temporal, serial assessment of CoLab-score over time
requires serial (daily) measurements. Possibly the CoLab score for dynamic
assessment requires adaptation. We propose an observational cohort study, in
which we compare the serial changes over time in CoLab-score of COVID-19
patients who are mechanically ventilated at the ICU or are in isolation at the
nursing ward of Pulmonary Diseases. The algorithm will be determined daily.
Next to the laboratory parameters involved in the CoLab-algorithm, additional
routinely available laboratory parameters will be investigated for a possible
role to adapt the algorithm (e.g., by including objective numeric results of
the SARS-CoV-2 RT-PCR (Ct-values), SARS-CoV-2 viability PCR as well as
hematological characteristics of the immunocompetent peripheral blood cells).

Primary aim:
1. Can the (adapted) CoLab-score be used to determine the moment that a
patient with COVID-19 has overcome the infection and can be
transferred safely to a non-COVID ICU or ward?

Secundary aims:
2. What is the development over time of the (adapted) CoLab-score in COVID-19
positive patients in isolation at the ICU/ nursing ward Pulmonary
Diseases?
3. After how many days does the (adapted) CoLab score show a normalization in
COVID-19 patients in isolation at the ICU/nursing ward
Pulmonary Disease
4. After how many days will the SARS-CoV-2 viability-PCR become negative in
comparison with the SARS-CoV-2 RT-PCR?
5. Is the evolution of the CoLab-score over time associated with the
development of multi-organ failure (serial organ failure assessment; SOFA-
score) over time in mechanically ventilated COVID-19 patients at the
ICU?
6. Is a low CoLab-score over time associated with negative SARS-CoV-2 viability
PCR testing in COVID-19 patients?
7. What is the impact (in costs / benefits and stakeholder experiences) of the
implementation of the CoLab score in standard practice?

DESIGN
Three cohorts of patients will be used: two existing (prospective and
retrospective) cohorts and a new prospective (third part).
First part: This will consist of the data of the Maastricht Intensive Care
Covid (MaastrICCht) cohort study. The design of this study is described
extensively elsewhere.(17) Briefly, this prospective cohort study was conducted
in patients admitted to the Intensive Care of the Maastricht University Medical
Centre+ (Maastricht UMC+), a tertiary care university teaching hospital. This
study cohort included al participants with respiratory insufficiency requiring
mechanical ventilation and at least one PCR positive for SARS-CoV-2 and/or a
chest CT strongly suggestive for COVID-19, based on a CORADS-score of 4-5
scored by a radiologist. Participants were followed until primary outcome was
reached (i.e., either death in the ICU or discharge from the ICU). During their
stay at the ICU, these patients had daily laboratory checks, including the
parameters necessary for running the CoLab-algorithm. These data are already
available and collected (n= 250 mechanically ventilated patients) and will be
used to verify the CoLab-algorithm and determine the cut-off of the CoLab
linear predictor (outcome) in predicting the moment that the patient had
conquer COVID-19 and investigate the association between CoLab-score and
SOFA-score over time.
Second part: In this retrospective part, we want to use data from serial
laboratory measurements of the COVID-19 positive patients that were admitted to
the ICU/nursing ward Pulmonary Diseasess of the other participating hospitals
in the period March 1st, 2020 and May 1st, 2021. With the use of these data,
the CoLab-algorithm with the new established cut-off for the linear predictor
(outcome) will be verified for accuracy in external derived datasets.
Third part: In this prospective part, we want to collect new serial laboratory
data to determine the CoLab score in COVID-19 positive patients admitted to the
ICU/nursing ward Pulmonary Diseases of Zuyderland MC in combination with
nasopharyngeal swabs/sputum for the SARS-CoV-2 viability PCR to proof that the
patients with a normalized CoLab-algorithm are really non-infectious anymore.

Nature and extent of the burden are minimal and risks are negligible, because
the load for the person is minimal (there is already a daily laboratory-blood
monitoring. The tests necessary to calculate the CoLab-score will be determined
in the left-over blood samples of that day. The PCR-tests will be performed in
sputum samples collected out of the tube of the mechanical ventilation