Incidence of thrombotic outcomes for patients hospitalized and discharged after COVID-19 infection.

To the editor,

Many publications reported an increased incidence of clinically significant venous and arterial thrombosis linked to COVID-19. Retrospective studies demonstrated an elevated incidence of venous thromboembolic complications with an incidence of 20,6–49% in intensive care unit populations [[1], [2], [3], [4], [5]]. Others have reported an incidence of arterial thrombotic complications of 3,6–3,7% in COVID-19 hospitalized patients [4,5].

This data has fueled some groups to propose the use of intermediate-dose thromboprophylaxis in patients at higher risk, especially patients with body mass index >30 or admitted to the ICU with a severe inflammatory syndrome or requiring mechanical ventilation [6,7]. Some groups have also proposed post-discharge prophylaxis even though randomized trials in the post-medical hospitalization setting have been deceiving.

We have evaluated the rates of in-hospital thrombotic events as well as rates of thrombosis after discharge.

We first did a retrospective cohort study of every adult patient hospitalized with a confirmed COVID-19 infection at the Centre Hospitalier Universitaire de Montréal from March 2020 until June 27th 2020. Most patients were admitted to a dedicated COVID medical ward for treatment and care, from which a subpopulation was transferred to the ICU for respiratory or hemodynamic support. Another set of patients, composed of patients coming from different hospitals after an acute COVID-19 infection, and were waiting for either negative tests to return to their care facility or needed rehabilitation or geriatric assessment, were admitted to a dedicated COVID rehabilitation unit.

We excluded patients who were hospitalized for conditions judged unrelated to COVID-19 infection as well of those who were in palliative care at the beginning of their admission. COVID-19 cases received systematic low-dose thromboprophylaxis although our intensive care unit (ICU) protocol allowed for intermediate-dose prophylaxis and full anti-coagulation in selected high-risk as of April 7th 2020 based on emerging literature on this subject. The dose increase was suggested for patients with BMI > 30, for ICU patients and those with a personal history of VTE not currently on anticoagulation.

A total of 552 patients with a positive COVID-19 test were hospitalized in our center. From these, 454 patients were eligible for evaluation as defined above. For the post-discharge assessment, a cohort 175 patients consenting patients could be contacted for further evaluation.

Of the 454 patients hospitalized with COVID-19 infection, 285 were admitted exclusively on the ward and another 78 were admitted in the ICU at some point in their stay. A distinct group of 91 patients were admitted to a COVID-19 rehabilitation ward from various institutions in the city. Thromboprophylaxis was administered to 91,2% of the ward population, 96,2% of the ICU population and 88% of the rehabilitation ward population. Failure to receive prophylaxis was related to high bleeding risk in most cases or dying within a very short period.

Patients characteristics and in-hospital thrombosis outcomes are described in Table 1 .

Table 1

Baseline characteristics and thrombotic outcomes in the hospitalized populations.

	Patients admitted exclusively on the ward(N = 285)	Patients admitted to the ICU (N = 78)	Patients admitted to the rehabilitation ward(N = 91)
Median age – yr [IQR]	72.0 [55–84]	63 [50–68]	86 [79–91]
Males – n (%)	146 (51)	49 (63)	31 (34)
Residency prior to admission
Home – n (%)	152 (53)	65 (83)	20 (22)
Senior homes – n (%)	77 (27)	5 (6.4)	45 (49)
Long-term care facility – n (%)	56 (20)	8 (10.3)	26 (29)
Length of stay – median days [IQR]	10 [5–22]	18 [10–43]	25 [16–36]
Length of stay in ICU – median days [IQR]	–	10 [5–26]	–
Median weight – kg [IQR]	71,9 [60–87](N = 200)	82,6 [66–89](N = 73)	66,0 [46–77](N = 49)
Comorbidities
Smoking
Never – n (%)	223 (78,2)	64 (82,1)	73 (80,2)
Past or active – n (%)	62 (21,8)	14 (7,9)	18 (19,8)
COPD – n (%)	43 (15)	7 (9.0)	18 (20)
CAD – n (%)	62 (22)	13 (17)	31 (34)
Hypertension – n (%)	157 (55)	33 (44)	57 (63)
Dyslipidemia – n (%)	106 (37)	22 (29)	38 (42)
Diabetes – n (%)	88 (31)	21 (27)	30 (33)
Stroke – n (%)	31 (11)	4 (5.3)	13 (14)
Atrial fibrillation – n (%)	46 (16)	6 (7.7)	23 (25)
Past history of VTE – n (%)	16 (5.6)	7 (9.0)	0 (0)
Active cancer – n (%)	33 (12)	7 (9.0)	6 (7.0)
Immunosuppression – n (%)	23 (8.0)	7 (9.0)	4 (4.4)
Supportive care
Oxygen supplementation – n (%)	127 (45)	72 (92)	12 (13)
Mechanical invasive ventilation – n (%)	0 (0)	50 (64)	0 (0)
Renal replacement therapy – n (%)	5 (2)	17 (22)	0 (0)
Extracorporeal membrane oxygenation – n (%)	0 (0)	4 (5,1)	0 (0)
Mortality – n (%)	61 (21)	30 (38)	6 (7)
Thromboprophylaxis at hospital
None	28 (9.8)	2 (2.7)	11 (12)
Low dose	200 (70)	20 (26)	59 (65)
Intermediate	11 (3.9)	37 (47)	0 (0)
Therapeutic	43 (15)	16 (21)	21 (23)
Significant bleedinga – n (%)	3 (1.0)	8 (10)	1 (1.0)
Thrombotic outcomes
Arterial – N (%)	7 (2,5)	4 (5,1)	0 (0)
Stroke	6	3	0
Acute coronary syndrome	1	0	0
Others	0	1 (aortic thrombus)	0
Venous – N (%)	5 (1,8)	16 (20,5)	1 (1,1)
Pulmonary embolism	5	14	1
Deep vein thrombosis	0	0	0
Catheter-related thrombosis	0	2	0

Any major bleeding or significant non-major bleeding according to the ISTH definition.

After a median length-of-stay of 10 (IQR 5–22) days in the ward population, we found an incidence of 2,5% (7 events) and 1,8% (5 events) of arterial and venous thrombotic complications, respectively. For the arterial thromboses, six were strokes, with two as a presenting feature of the infection and the other was an acute coronary syndrome, also as a presenting feature. All VTEs were pulmonary embolisms.

In the ICU population, after a median stay of 18 (IQR 10–43) days, we found an incidence of 3,9% (three events) and 21% (16 events) of arterial and venous thrombotic complications. Of the arterial thromboses, two were strokes and the other one was an aortic thrombus, all of them present at admission. Most of the VTE complications were pulmonary embolisms (n = 14), whereas two events were catheter-related. We did not observe dialysis or ECMO circuit thromboses. Among this higher-risk population, we were not able to show an impact of thromboprophylaxis dose on the incidence of thrombotic events. Notably, 26% of our very high-risk ICU population presented a thromboembolic complication even if 68% of those were either treated with intermediate dose thromboprophylaxis or full anticoagulation therapy.

In the inpatient rehabilitation cohort, only one patient was found to have a pulmonary embolism, for an incidence of VTE of 1,1%.

As others have shown, we found a statistical difference between patients who had presented a thrombosis and those without thrombosis for D-dimers > 1500 μg/L (p < 0,0003), fibrinogen > 6 g/L (p < 0,0001), CRP (p = 0,047), LDH > 300 (p = 0,034) and elevated neutrophil values (p = 0,037).

In the outpatient cohort, a total of 175 patients consented to participate. From these, 140 patients were discharged from the medical ward and 35 from the rehabilitation ward. Median follow-up was 68,5 days and 60 days, respectively for discharged ward patients and rehabilitation patients. Characteristics and thrombotic outcomes of these patients are shown in Table 2 .

Table 2

Baseline characteristics and thrombotic outcomes in the outpatient populations.

	Patients discharged from the medical ward(n = 140)	Patients discharged from the rehabilitation ward(n = 35)
Median age – yr [IQR]	63 [48–75]	82 [75–89]
Males – n (%)	75 (54)	13 (37)
Residency at discharge
Home – n (%)	106 (76)	9 (26)
Retirement home – n (%)	19 (14)	17 (49)
Long-term care facility – n (%)	15 (11)	9 (26)
Length of stay – median days [IQR]	11,5 [6–21]	19 [13–28]
ICU stay – n (%)	24 (17)	0
Length of stay (ICU) – median days [IQR]	15 [6–23]	–
Comorbidities
Smoking
Never – n (%)	121 (86,4)	31 (88,6)
Past or active – n (%)	19 (13,6)	4 (11,4)
COPD – n (%)	11 (8)	4 (11,4)
CAD – n (%)	21 (15)	15 (43)
Hypertension – n (%)	65 (46)	26 (74)
Dyslipidemia – n (%)	45 (32)	16 (46)
Diabetes – n (%)	37 (27)	14 (40)
Stroke – n (%)	6 (4,3)	8 (23)
Atrial fibrillation – n (%)	17 (12)	9 (26)
Past history of VTE – n (%)	3 (2,1)	0 (0)
Cancer – n (%)	16 (11)	1 (2,9)
Immunosuppressed – n (%)	10 (7.1)	1 (2,9)
Anticoagulation at discharge – n (%)	14 (10)	10 (29)
Thromboprophylaxis at discharge – n (%)	0	0
Thrombotic outcomes
Total – n (%)	1 (0,71)	0 (0)
Venous – n (%)	1 (0)	0 (0)
Arterial – n (%)	0 (0)	0 (0)

Overall, only one patient discharged from the hospital ward presented a pulmonary embolism 9 days after discharge, an incidence of arterial or venous complication of 0% (95% CI 0–2,1%) and 0,71% (95% CI 0–2,1%). The incidence was 0% (95% CI 0–8,6%) in patients discharged from the rehabilitation ward.

Our results for hospitalized patients are in accordance with multiple other groups. In a cohort of 81 ICU patients from China, VTE was diagnosed in 25% of them, but thromboprophylaxis was not provided to most of these patients [8]. One Dutch study found a cumulative incidence of 49% venous thromboembolic events in 184 patients admitted to the ICU [4], whereas another Dutch group reported a cumulative incidence of 34% symptomatic venous thromboembolic events in a cohort of 74 ICU patients [3]. A French group reported 25% of thromboembolic complications in a COVID-19-related acute respiratory distress syndrome cohort, a much higher rate as compared to a non COVID-19 similar group [1]. One center in France reported an incidence of 40% arterial and venous thrombotic complications in a cohort of 92 patients [9]. An Italian study reported VTE in 7,7% of ward patients and a cumulative rate of 27,6% in ICU patients [5].

Arterial thrombosis of 2,5% in ward patients, 3,9% in ICU patients were in accordance with the incidence found in an Italian and Spanish cohort [5,10]. Of note, 60% of these events were presenting features, suggesting that arterial disease may be an early feature of COVID-19 infection, as reported by others.

We did not find a statistical association between the type of thromboprophylaxis and the incidence of thrombosis in the ICU population. Some 38% of patients that were fully anticoagulated presented a thrombotic event, a higher proportion than with standard and intermediate dose thromboprophylaxis. We think this could represent an “impending doom” bias, with patients presenting with more worrisome clinical and laboratory features being treated more aggressively. There was also 10% of significant bleeding in that particular population. Similar findings were reported in a French study in which 21% of patients with COVID with full anticoagulation experienced a hemorrhagic complication [9]. Those results suggest that increases in anticoagulation doses should be studied in randomized studies before widespread adoption in this setting.

For patients who were discharged, we found an incidence 0,71% in the previously hospitalized in the medical ward and 0% in the rehabilitation ward populations. Globally, this data suggests that even if acutely infected COVID-19 patients have a high incidence of thromboembolic complications, the risk diminishes quickly after the acute phase. Those results are in accordance with an abstract presented at the 2020 virtual ISTH Congress, in which 102 patients, including 26 who were hospitalized in an intensive care unit, presented an incidence of less than 1% VTE (one asymptomatic deep vein thrombosis).

In this report, we confirm a low thromboembolic risk in discharged patients after a COVID-19 infection. These data suggest that this population should not systematically require prolonged thromboprophylaxis until prospective randomized trials are performed in this setting.

Funding

This work is supported by the Défi Respire foundation (www.icm.qc.ca).

Disclosure

All the authors have nothing to disclose.

Declaration of competing interest

No conflict to declare.