Dynamic Blood Tests Reveal Predicted Week of Death And Tricky Week For Surveillance In Severely Ill COVID-19 Patients

Background: Blood laboratory tests are the most reliable methods for the diagnosis and assessment of vital organs’ functions and the body’s response to infection. Herein, we compared the results of dynamic blood tests between the survivor and non-survivor group of patients with coronavirus disease 2019 (COVID-19) and aimed to determine the predicted and tricky week for death and surveillance. Methods: The survivor and non-survivor groups were compared using biochemical blood tests, routine blood tests, and coagulation blood tests over four weeks of investigation. Results: Blood urea nitrogen, creatinine, high-sensitivity C-reactive protein, total bile acid, neutrophil count, white blood cell count, D-dimer, brin and brinogen degradation product, and prothrombin time showed signicantly higher levels in the non-survivor group than the survivor group. Only pre-albumin, eosinophil count, lymphocyte count, red blood cell count, platelet count, hemoglobin, and prothrombin activity tests were signicantly higher in the survivor group than the non-survivor group. Generally, the third week of the non-survivor’s group could be regarded as the predicted week for death based on all tests except for creatinine, pre-albumin, total bile acid, monocyte count, white blood cell count, and prothrombin activity. The tricky week in the non-survivor group was the second week in all tests except for pre-albumin, basophil count, eosinophil count, lymphocyte count, platelet count, D-dimer, and brin and brinogen degradation product. Conclusions: Based on our study, specic attention should be given to some weeks with respect to their related tests as predicted or tricky for death or surveillance, respectively.


Background
In December 2019, the novel coronavirus disease 2019 (COVID-19) was rst identi ed in Wuhan, China. The outbreak was eventually declared an international public health emergency in January 2020 and later a pandemic in March 2020 by the World Health Organization [1]. By October 2020, there were > 75 million con rmed cases of COVID-19 globally and > 1.6 million deaths [2]. COVID-19 infection is caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). This virus belongs to the Coronaviridae family and partially to SARS-CoV-2 and middle east respiratory syndrome-related coronavirus (MERS-CoV), that have caused previous epidemics in China and the Middle East, respectively [3].
Although nearly 80% of patients with COVID-19 show mild or moderate symptoms, those with severe and critical infection are more likely to be admitted to the intensive care unit (ICU) with worse prognosis [4]. The clinical and laboratory ndings of COVID-19 patients may predict the severity in some cases. Elevated liver enzymes occur in a median of 15% [5] and up to 58% [6] of patients with COVID-19, and kidney damage was not uncommon [7]. Around 40% of hospitalized patients with COVID-19 in China were at a high risk of developing venous thromboembolism [8], had abnormal coagulation factors, and were associated with poor prognosis [9].
Although several laboratory ndings on admission have been reported [10], the dynamic changes during the hospitalization period for both COVID-19 survivors and non-survivors have rarely been described. In addition, at speci c time periods, some blood tests showed normal or abnormal records but without any previous correlation to surveillance or death. In the present study, we investigated three groups of laboratory ndings and their outcomes on surveillance and death with determination of speci c weeks during treatment/admission. As the pandemic evolves, our ndings may provide a scienti c basis for death expectancy during four weeks of hospitalization for patients who survived or died during this this period.

Study design and participants
This retrospective observational study was conducted at Huanggang Central Hospital, Wuhan, China. In this study, COVID-19 was con rmed by real-time reverse-transcription polymerase chain reaction-based detection of SARS-CoV-2, and patients severely ill with COVID-19 were eligible for inclusion. The severity of COVID-19 was classi ed based on the Guidance for Coronavirus Disease 2019 (6th edition) released by the National Health Commission of China [11]. Brie y, the criteria for severe cases included dyspnea, respiratory rate ≥ 30 breaths per minute, blood oxygen saturation ≤ 93%, partial pressure of arterial oxygen to fraction of inspired oxygen ratio ≤ 300, and/or lung in ltrates > 50% within 24-48 hour.
Severely ill COVID-19 patients were categorized into two groups: survivors (S) and non-survivors (NS) according to the nal clinical outcomes of surveillance or death. Each group had 21 patients, and the laboratory tests were investigated over one month starting from the day of admission until the fourth week of hospitalization.
These tests were categorized into three groups; biochemical blood tests, routine blood tests, and coagulation blood tests (Table 1).

Results
Different records between weeks in the survivor and non-survivor groups

Biochemical blood tests
The biochemical pro le was represented by 18 tests. The non-signi cant values between/within survivors and non-survivors are shown in supplementary gure 1. Blood urea nitrogen, creatinine, high-sensitivity C-reactive protein (CRP), pre-albumin, and total bile acid tests showed several signi cant differences within and between each group ( Figure 1).

Blood urea nitrogen
In the blood urea nitrogen test, no signi cant difference was found within the survivor group, while several differences were reported within the non-survivor group and between both the survivor and non-survivor groups.

Pre-albumin
In the survivors' group, pre-albumin was signi cantly high in week 4S compared to week 2S (P≤0.05). On the other hand, a signi cant increase was found in week 4S in comparison to week 1NS (P≤0.01), week 2NS (P≤0.001), and week 4NS (P≤0.05) (Fig. 1D).

Total bile acid
Within the survivors' group, no signi cant difference was found. The highest value was recorded in week 4NS with highly signi cant differences in relation to week 3NS, week 2NS, week 1NS, week 4S, week 3S, week 2S, and week 1S (P≤0.001) (Fig. 1E).

Routine blood tests
The routine blood tests were represented by nine tests. All tests showed several signi cant inter-and intragroup differences in each group (Fig. 2).

Basophil count
The highest value of basophil cells was recorded in week 1S and week 4S. Signi cant differences were found between week 4S and week 2S (P≤0.05) and between week 2S and week 1S (P≤0.01) ( Fig. 2A).

Eosinophil count
In the survivors' group, a signi cantly higher value was found in week 4S than week 2S (P≤0.01). However, week 4S showed higher values than week 2NS, week 3NS (P≤0.01), and week 1NS in the non-survivors group (P≤0.05) (Fig. 2B).

Lymphocyte count
Signi cant differences were found within and between the survivor and non-survivor groups. The highest record was found in week 1S. Week 2S was signi cantly higher than week 3NS and week 4NS (P≤0.01), while week 3S was signi cantly higher than week 1NS (P≤0.01), week 2NS, week 3NS, and week 4NS (P≤0.001).

Monocyte count
The highest monocyte count was found in week 1S and week 1NS, while signi cant differences were found only between week 1NS and week 2S (P≤0.05) (Fig. 2D).

Neutrophil count
The neutrophil count was higher in the non-survivor group especially in week 1NS, week 3NS, and week 4NS week 2NS and between week 3NS and week 4NS (P≤0.05) (Fig. 2I).

Coagulation blood tests
Seven blood coagulation tests were dynamically tested over the four weeks among both survivors and nonsurvivors. Activated partial thromboplastin time, brinogen, and thrombin time did not show any signi cant difference between and within the survivors and non-survivors (Supp. Fig. 2). On the other hand, D-dimer, brin and brinogen degradation products, prothrombin activity, and prothrombin time showed signi cant differences within and between groups (Fig. 3).

Predicted or tricky weeks for surveillance and death
Predicted or tricky weeks as blood test biomarkers for surveillance and death have been recorded (Table 2). In our results and discussion, we will be focusing on the early predicted weeks for death (week 1NS, week 2NS, or week 3NS) and accordingly, the tricky weeks for surveillance for the clinicians to consider them as risk factors with time speci cation related to their underlining condition.

Biochemical blood tests
In the non-survivor group, the predicted weeks for death were likely week 3NS with respect to both blood urea nitrogen and high-sensitivity CRP, while for creatinine and pre-albumin, week 1NS and both week 1NS and week 2NS, respectively, were the predicted weeks for death. Accordingly, the tricky weeks for surveillance in blood urea nitrogen and high-sensitivity CRP were weeks 1NS and 2NS. Both week 2NS and week 3NS could be regarded as tricky weeks for surveillance in creatinine, while only week 3NS in the pre-albumin test may be the tricky week. All the investigated weeks for total bile acid in the non-survivor group could be regarded as tricky weeks for surveillance ( Table 2).

Routine blood tests
Within the non-survivor group, all weeks for basophil count and lymphocyte counts could be regarded as predicted weeks for death, while for eosinophil count and hemoglobin test, only the rst three weeks and the rst and third weeks, respectively, could be regarded as the predicted weeks for death. Neutrophil count, platelet count, and red blood cell count shared the 3NS as predicted week for death, while the additional week for prediction was week 2NS and 1NS for platelet count and red blood cell count, respectively. On the other hand, the early tricky week for surveillance was week 2NS for the hemoglobin test and red blood cell count, all the investigated weeks for monocyte count, the rst two weeks for neutrophil count, the rst three weeks for white blood cell count, and only week 1NS and 2NS for platelet count and red blood cell count, respectively (Table 2).

Coagulation blood tests
For D-dimer and brin and brinogen degradation product tests, all the weeks in the non-survivor group could be the predicted weeks for death, while only week 3NS may be the predicted week for death in prothrombin time. Conversely, week 1NS, week 2NS, and week 3NS can be regarded as the tricky weeks for surveillance in prothrombin activity, while only week 1NS and week 2NS can be the tricky weeks for surveillance in prothrombin time (Table 2). Other studies showed that blood urea nitrogen was regarded as one of the predictive tests for surveillance with signi cant elevation in the non-survivor group [12,13]. In addition to blood urea nitrogen, creatinine was not found to be different between severe and moderate cases of COVID-19 [13] when measured during the rst weeks of infection. Our results showed that creatinine was obviously and signi cantly higher in the fourth week of the non-survivor group than week 2NS and week 1S, week 2S, week 3S, and week 4S.

Discussion
In in ammation, high-sensitivity CRP is regarded as one of the most common and accessible tests. In the third and fourth week in the non-survivor group, high-sensitivity CRP was obviously elevated when compared to others. Wang et al. found that even in the early stages of severe and critical COVID-19, CRP was higher than the normal level and positively correlated to the lung lesion [14] but without declination in critically ill patients in the latter period of hospitalization [15]. Along with that, pre-albumin is regarded as the precursor of different Basophil count was reported to be normal and relatively higher at the early and late stage of infection in the survivor group than all weeks in the non-survivor group. A prediction of severity was concluded with a lower than normal level at the early stage of infection and even in comparison with the control group [19].
Eosinopenia (< 0.02×10 /L) which has been reported at the admission time of fatal cases of COVID-19, was regarded as a risk factor [20]. We reported a normal eosinophil count at week 1S and despite its declination during week 1S and week 2S, elevation was noticed at week 3S promising a good prognosis. Liu et al. noticed that the antiviral treatment of COVID-19 patients had improved eosinopenia at the late stage before discharge [21]. Normal lymphocyte levels at the week of admission were regarded as a good prognostic marker for the following weeks [22]. A normal level of lymphocyte count at week 1S was reported despite its declination at only week 2S, while all four weeks in the non-survivor group recorded less than the normal level. In the survivors group, the monocyte count was within the normal level during the hospitalization period, while it was on its highest normal level at week 1NS suggesting a bad prognosis or severe disease [23]. In addition, for neutrophil count, the survivors group showed normal levels throughout the four weeks, while week 1NS showed a signi cantly higher level than week 2S, week 3S, and week 4S. This high value decreased in week 2NS but again elevated in week 3NS and week 4NS to be signi cantly higher than other weeks in the survivors and nonsurvivor groups. The comparison followed the same pattern of signi cantly higher values of neutrophil count in the non-survivors than survivors but in general and without any week speci city [24]. The latter study also showed a signi cantly high level of white blood cell count in the non-survivors than survivors [24]. Despite that, all the weeks of both groups showed normal level of white blood cell count except week 4NS, which showed a signi cantly high level in comparison to all other weeks.
Among the red blood cell and platelet counts and hemoglobin level, we found that red blood cell count during week 1S was within the normal range but signi cantly higher than week 1NS. This meant that a slight abnormal level in week 1NS may be a predictive risk factor for COVID-19 patients despite the elevation in week 2NS; however, it returned to lower than normal levels in week 3NS and week 4NS. Yuan et al. found that red blood cell count was signi cantly lower in critically ill COVID-19 patients than in regular COVID-19 patients, con rming that the red blood cell count may play an important prognostic marker for worse prognosis due to insu cient oxygen transportation [25]. In relation to the impact of cytokine storm to thrombocytopenia and low platelet count [26], the results of Mehta et al. were in accordance with ours, in that the decline in platelet count started at week 1NS followed by week 2NS/3NS and week 4NS. For hemoglobin, other studies did not nd signi cant differences between survivors and non-survivors [24,27], while our study provide an insight from week 1S and week 1NS that either surveillance or death are expected, respectively, as the week 1S value was signi cantly higher than that of week 1NS. Coagulation blood tests D-dimer has extensively been studied for its relationship to COVID-19 infection and as a biomarker for disease severity [28][29][30][31], especially in the post admission stage [32]. In the current study, all weeks in both the survivor and non-survivor groups showed an elevated level of D-dimer, but signi cantly higher values were seen in week 3NS and week 4NS compared to other weeks. Fibrin and brinogen degradation products during the whole period of hospitalization in the survivor group were within normal levels, while they were signi cantly elevated and exceeded normal levels in the non-survivor group in all weeks of hospitalization.
A meta-analysis was performed on 17 studies to determine the relation of high abnormal levels of brin and brinogen degradation products at admission to poor outcome [33]. For prothrombin activity, no obvious change was observed for all the weeks in the survivor group, as it ranged between 111% and 114%. The level was within normal levels in week 1NS, week 2NS, and week 3NS, but in week 4NS, it declined to below the lowest normal level. In accordance with this nding, Luo et al. reported a signi cantly low level of prothrombin activity in the non-survivors group [34]. Although prothrombin time was recorded to be normal and steady on 11 second, it showed a gradual increase in the non-survivors' group starting from week 2NS, revealing it as a poor prognostic marker. A signi cant prolongation in prothrombin time was recorded in non-survivors compared to survivors [9,34].
Predicted or tricky weeks for surveillance and death

Biochemical blood tests
For the biochemical tests, we reported several predicted weeks for death. In our study, week 3NS could be a predicted week for death in blood urea nitrogen test. In another study and from the third week, the survival curve declined to 50% when the blood urea nitrogen exceeded 4.6 mmol/L with high speci city, sensitivity, and a strong predictive value for mortality [35]. A signi cantly high value of creatinine at admission was reported with a progressive increase in the non-survivors group until death [36], revealing that early attention is important for this test. High-sensitivity CRP was recorded to be signi cantly higher at the week of death compared to week of admission, but in this study, the time-interval between admission and death was only 12 days [37]. Lower than normal platelet counts started from week 2NS and later, which was clearly indicated by Zhao et al.
to be a signi cant biomarker for poor prognosis and death [45]. Our study reported lower than the normal level for the red blood cell count in the rst and third week in non-survivor group expecting an early predicted week for death. Yuan et al. also reported a signi cant lower red blood cell count in severe and critically ill than in the regular group but without speci cation to which week the record was reported [46].
Some blood tests should be considered as tricky tests in speci c time periods. Week 2NS may exhibit normal hemoglobin levels, but it was a poor prognostic marker as indicated by Lippi and Mattiuzzi in which the continuous decrease in hemoglobin is a risk factor for death [44]. Zhao et al. reported a signi cant increase in neutrophil count at the admission ( rst week) in non-survivors compared to survivors, which was the same as in our study. Despite this similarity, in both ours and Zhao et al's studies [45], the rst week of the non-survivor group demonstrated that the neutrophil count was within the normal level, taking into consideration that this normal level may still be a tricky sign or tricky week for surveillance. The same explanation may be valid for the platelet count [45]. The red blood cell count in our study was within the normal level in the second week of nonsurvivors group taking into consideration that this week could be a tricky week for surveillance. To our knowledge, no previous study has speci ed this trickiness before. The rst three weeks in the non-survivor group could be the tricky weeks for death for the white blood cell count as it is within the normal rage.  [47].

Coagulation blood tests
In the non-survivor group and for D-dimer test and brin and brinogen degradation products test, we found that week 1NS to week 4NS were the predicted weeks for death. Accordingly, studies recorded a signi cantly high level of D-dimer [48] and brin and brinogen degradation products [34] correlating them to poor prognosis in severe COVID-19 patients. We recorded several signi cant increases in prothrombin time in week 3NS compared to other weeks, while Tang et al. recorded a signi cantly high correlation of prothrombin time to mortality on the 28 th day [49].
On the other hand, the rst two weeks and the rst three weeks in the non-survivor group were regarded as tricky weeks for surveillance in prothrombin time and prothrombin activity tests, respectively. Similar to our results, Luo et al. predicted that the prothrombin time at admission was a predictor for mortality but still at the normal level [34], while Tang et al. reported abnormally high prothrombin time at admission and within the whole period of hospitalization in the non-survivors group [9]. Week 1NS, week 2NS, and week 3NS were found to be the tricky week for surveillance in prothrombin activity, which has been also reported that by another study that this test was related to mortality in non-survivors despite its normal level [50].

Conclusion
Dynamic blood tests in COVID-19 patients reveal important indicators of the physiological function of vital organs, in addition to providing a focused attention on a speci c time period during hospitalization. The signi cant differences between weeks in survivors and non-survivors, and whether they are within or outside the normal range, could be helpful in another interpretation for surveillance or death as predicted or tricky weeks.