After Wuhan, in China, Italy has become the European country with the highest number
of cases and deaths by Sars-Cov-2.[1]
[2]
It has been reported from China the presence of altered coagulation parameters in
a significant proportion of hospitalized patients.[3]
[4]
[5] In addition, increased D-dimer (DD) levels have gained particular attention as predictors
of acute respiratory distress syndrome development, the need of intensive care unit(ICU)
admission, or death.[6]
[7] It is well known that DD is a highly nonspecific marker which may be only the mirror
of the clotting activation following the inflammatory process.[8]
[9] Nevertheless, several thrombotic complications have been reported in patients with
Covid-19 treated in ICU and the possible treatment with heparin appears to be very
attractive.[10]
Before this therapeutic application, it is mandatory to verify and report to the scientific
community which is the trend of coagulation parameters, so that we can accumulate
several observations useful to design ad hoc interventional trials.
In this scenario, we decided to analyze the routine coagulation parameters—prothrombin
time (PT), activated partial thromboplastin time (aPTT), DD, and platelet count—of
a consecutive series of patients with documented infection by Sars-Cov-2 hospitalized
in Careggi Hospital, Florence, Italy.
Methods
Two hundred and nine consecutive patients (133 males/76 females) with confirmed Sars-Cov-2
infection (according to the World Health Organization criteria), admitted to the Careggi
Hospital, Florence, from March 2 to April 7, 2020, were enrolled. Exclusion criterion
was the direct admission in ICU.
The clinical outcomes were monitored up to April 7, 2020 (median follow-up: 11 days
[interquartile range [IQR]: 7–14]).
The results were given as mean ± standard deviation or median (IQR), wherever appropriate.
Distributed quantitative variables were compared using the Student's t-test and the Mann–Whitney U test, as appropriate. A receiver operating characteristic (ROC) curve analysis was
used to determine the ability of the analyzed parameters to distinguish between patients
with different endpoints: death or hospital discharge. The optimal cut-off point was
calculated by determining the level that provided the greatest sum of sensitivity
and specificity.
Results
Thirty-one patients died (14.8%) and 117 patients had been discharged (56%). Fifty-one
patients (24.4%) have been transferred to an ICU. Eight thrombotic events were documented
(3.8%) in symptomatic patients, in the absence of routine compression ultrasound execution.
Laboratory values in the total population are reported in [Table 1].
Table 1
Demographic and laboratory parameters of enrolled Sars-Cov-2 patients
|
Variables
|
All (n = 209)
|
Alive (n = 178)
|
Dead (n = 31)
|
p-Values
|
Discharged (n = 117)
|
Nondischarged (n = 92)
|
p-Values
|
|
Age, mean ± SD
|
65.9 ± 14.6
|
63.5 ± 14.0
|
79.8 ± 9.8
|
< 0.001
|
59.9 ± 14.0
|
73.6 ± 11.6
|
< 0.001
|
|
Sex, M N (%)
|
133 (63.6)
|
112 (62.9)
|
21 (67.7)
|
0.607
|
74 (63.2)
|
59 (64.1)
|
0.895
|
|
PT, s
|
14.0 ± 2.5
|
13.8 ± 2.4
|
15.2 ± 3.1
|
0.003
|
13.4 ± 1.1
|
14.8 ± 3.5
|
< 0.001
|
|
INR
|
1.23 ± 0.21
|
1.22 ± 0.21
|
1.31 ± 0.21
|
0.002
|
1.18 ± 0.10
|
1.30 ± 0.28
|
< 0.001
|
|
aPTT ratio
|
1.02 ± 0.14
|
1.02 ± 0.14
|
1.02 ± 0.15
|
0.858
|
1.00 ± 0.09
|
1.03 ± 0.19
|
0.786
|
|
Platelet count, ×109/L
|
187 (145–231)
|
189 (149–233)
|
174 (111–204)
|
0.127
|
191 (156–235)
|
183 (141–217)
|
0.088
|
|
MPV, fL
|
10.6 (10.1–11.4)
|
10.5 (10.0–11.3)
|
11.4 (10.6–12.4)
|
< 0.001
|
10.4 (9.9–11.1)
|
10.9 (10.5–11.9)
|
0.001
|
|
D-dimer, ng/mL (n = 135)
|
794 (513–1,247)
|
739 (498–1,160)
|
1,149 (757–2,049)
|
0.002
|
669 (490–960)
|
969 (550–1,670)
|
0.011
|
Abbreviations: aPTT, activated partial thromboplastin time; M, males; INR, international
normalized ratio; MPV, mean platelet volume; PT, prothrombin time; SD, standard deviation.
Note: Statistical significant values (p < 0.05) are depicted in bold.
We compared coagulation parameters at hospital admission between survivors versus
nonsurvivors.
Nonsurvivor patients were older and, at hospital admission, they had significantly
higher PT-international normalized ratio (INR), PT seconds, and mean platelet volume
(MPV) values with respect to survivors ([Table 1]). No differences were detected in aPTT ratio and platelet count. ROC curves revealed
that the cut-offs with the highest sensitivity and specificity for the endpoint “death”
were: 1.40 for PT-INR, 1,000 ng/mL for DD, and 11.9 fL for MPV. By defining “Sars-Cov-2 induced coagulopathy” as a combination of the three parameters (i.e., the presence of a value EQUAL or
HIGHER than the reported cut-off for more than one parameter), area under the curve
(AUC) increased to 0.78 ± 0.06 (p < 0.05; [Fig. 1A]). The addition of the so-defined “Sars-Cov-2 induced coagulopathy,” to a model that included age, which is “per se” a strong predictor of death (AUC
0.89 ± 0.04), slightly increased the AUC (0.92 ± 0.03; p = 0.10) for the detection of death ([Fig. 1C]). At multivariate regression analysis, adjusted for age and sex, age and Sars-Cov2-induced
coagulopathy were independent predictors of death (age: odds ratio [OR] = 1.17, 95%
confidence interval [CI] 1.07–1.27; p < 0.0001/Sars-Cov2-induced coagulopathy: OR = 2.72, 95% CI 1.20–6.17; p = 0.016).
Fig. 1 Receiver operating characteristic (ROC) curves according to death and hospital discharged.
ROC curves of PT-International Normalized Ratio (INR), D-dimer (DD) and Mean Platelet
volume (MPV) according to death (1-A) and hospital discharge (1-B), when the three parameters were above (1-A) or below (1-B) the cut-off values. ROC curves for death of the two models of logistic
regression: model 1 (age) or model 2 (age + combined coagulative parameters*) (1-C). ROC curves for hospital discharge of the two models of logistic regression: model
1 (age) or model 2 (age + combined coagulative parameters*) (1-D).
In addition, we compared coagulation parameters at hospital admission between discharged
versus nondischarged patients. Discharged patients were younger and, at admission,
had significantly lower PT-INR, PT seconds, DD, and MPV values with respect to the
others ([Table 1]). No differences were detected in aPTT ratio and platelet count. ROC curves revealed
that the cut-offs with the highest sensitivity and specificity for the endpoint “hospital
discharge” were: 1.23 for PT-INR, 1,000 ng/mL for DD, and 10.6 fL for MPV. By defining
“NO Sars-Cov-2 induced coagulopathy” as a combination of the three parameters (i.e., the presence of a value LOWER than
the reported cut-off for more than one parameter), AUC increased to 0.79 ± 0.04 (p < 0.05; [Fig. 1B]). By adding the “NO Sars-COv-2 induced coagulopathy” to a model that included age, which is “per se” a strong predictor of hospital discharge
(AUC 0.80 ± 0.04), significantly increased the AUC (0.88 ± 0.03; p = 0.016) for the detection of hospital discharge ([Fig. 1D]).
Discussion
These results demonstrate that patients at hospital admission for Sars-Cov-2 infection
often have an alteration in the routine coagulation parameters. In particular, PT-INR,
DD, and MPV are the three parameters which, together with age, resulted to be significantly
associated with death and discharge, so allowing us to identify two clusters of inpatients
with worse or better prognosis.
In particular, the absence, at hospital admission, of an involvement in clotting and
fibrinolytic pathways identifies a group of patients at low risk of complications.
We think that the most relevant result is that this score not only maintains but increases
the AUC for death or discharge if associated with age, which is “per se” the strongest
clinical predictor of outcome.
It is well known that thrombosis is associated with inflammation and vice versa.[11]
[12] In the model of sepsis, it has just been validated as score (SIC – sepsis-induced
coagulopathy score) in which PT-INR, fibrinogen, and platelet count are associated
with a worse prognosis. In the setting of Sars-Cov-2 infection, it has been demonstrated
that both clotting activation and fibrinolysis are crucial.[13]
[14]
[15]
[16]
[17]
[18] Indeed, our data demonstrate that also DD levels help us to identify patients at
higher or lower risk. In addition, in our group, platelet size and not platelet count
is significantly associated with clinical outcomes: we know that MPV is a marker of
platelet function as it is positively associated with indicators of platelet activity,
including aggregation and release of thromboxane A2, platelet factor 4, and b-thromboglobulin.[19]
[20]
A subsequent step might be a clinical trial designed to randomize patients with “Sars-Cov-2 coagulopathy” to an intensive antithrombotic treatment, to verify a possible impact on prognosis.
Possible intervention in this area might be heparin treatment at intermediate or therapeutic
dosage or other treatment able to interfere with the inflammatory process as suggested
by the study of Tang et al.[6]
A limitation of this study is the limited number of patients recruited from a single
center; in addition, some patients are still hospitalized at the time of manuscript
submission.
Nonetheless, these data suggest a possible clinical utility of a score based on routine
coagulation parameters to stratify the prognosis of hospitalized patients with Sars-Cov-2
infection.
