Increased anticoagulation reduces proximal deep vein thrombosis in mechanically ventilated COVID-19 patients: Venous thrombosis prevention & COVID-19.

Dear Editor,

Coronavirus disease-2019 (COVID-19) has been associated with life-threatening thromboembolic complications due to increased inflammation, marked hypercoagulability and endothelial activation.1, 2, 3 Preventing deep vein thrombosis (DVT) and pulmonary embolism is important, since ∼10% of COVID-19-related deaths are caused by pulmonary embolism complicating DVT. Recently, enhanced prophylactic or therapeutic anticoagulation dose regimens have been recommended by experts and adopted in some centres. , However, direct comparative studies of the different anticoagulation regimens are lacking.

In our intensive care unit (ICU), we established a DVT prevalence of 46% in mechanically ventilated COVID-19 patients on standard prophylactic anticoagulation and subsequently, we increased anticoagulation to reduce thromboembolic complications. We designed this before-after observational exploratory study to evaluate the risk/benefit ratio of increased (IA) versus standard prophylactic anticoagulation (SPA) in mechanically ventilated COVID-19 patients. The study was part of the COVID-ICU and French COVID-19 cohort registries and received approval from the ethics committee of our institution (N°, IDRCB, 2020-A00256-33; CPP, 11-20.20.02.04.68737).

We included all consecutive patients admitted for COVID-19-related pneumonia requiring tracheal intubation. We excluded patients on long-term therapeutic anticoagulation before ICU admission. To diagnose DVT, an initial ultrasound was routinely performed during the first week after intubation, and in DVT-free patients, a second ultrasound was performed ∼1week later by certified sonographers (SV/PB) according to guidelines. The study was composed of two periods, defined according to the type of anticoagulation received from intubation to the first ultrasound examination. In both groups, if DVT was diagnosed, therapeutic anticoagulation was initiated.

Patients admitted from 2020/03/11 to 2020/04/01 (SPA group) received prophylactic anticoagulation with subcutaneous enoxaparin 40 mg once daily or unfractionated heparin 15000IU/day if creatinine clearance <15 mL/min. Patients admitted from 2020/04/02 to 2020/10/12 (IA group) received either prophylactic double-dose enoxaparin 40 mg twice daily or therapeutic anticoagulation with either enoxaparin 1 mg/kg twice daily or unfractionated heparin to reach plasma anti-Xa activity of 0.3–0.6IU/mL. Supportive care included optimized mechanical ventilation, vasopressors, sedation and muscular paralysis according to guidelines. Dexamethasone, antiviral and other immunomodulatory drugs were administered according to the physicians in charge.

The efficacy endpoint was the prevalence of femoral/popliteal DVT, known to be strongly associated with pulmonary embolism. The efficacy endpoint was also compared between patients treated with double-dose prophylactic enoxaparin (0.4 mg twice daily) and patients treated with standard enoxaparin prophylaxis (0.4 mg once daily). The safety endpoint was the number of patients with at least one major bleeding (MB) defined according to guidelines, i.e. bleedings causing death, decreasing hemoglobin by ≥2 g/dL, requiring transfusion of ≥2 blood units or occurring in a critical organ.

Quantitative variables are expressed as medians [25th–75th percentiles] and categorical variables as percentages. Parameters were compared between SPA and IA patients using Mann-Whitney and Fisher's exact tests as appropriate. An exploratory generalized multilinear regression model was built to adjust for parameters significantly different between groups. P-values ≤0.05 were considered significant. Based on the 26% prevalence of femoral/popliteal DVT in the SPA group and a presumed reduction to <5% in the IA group, 42 patients/group were required for 95% confidence interval and 80% statistical power.

Ninety-three patients were included, 50 in the SPA and 43 in the IA group. Baseline characteristics did not differ significantly between the groups (Table 1 ). The initial ultrasound was performed 2 days [1-4] post-intubation. Time from ICU admission to the first ultrasound was 4days [2-6] in the SPA versus 5days [3-8] in the IA group, P = 0.03. In 37 of the femoral/popliteal DVT-free patients, a second ultrasound was performed 8days [7-10] post-intubation. Anticoagulant treatment is presented in Table 1. At the time of the initial ultrasound C-reactive protein, fibrinogen and D-dimer were remarkably elevated at 223 mg/L [132-307], 7.6 g/L [6.2–8.6] and 3180 ng/mL [1495-6808], respectively. Twenty-nine patients (31%) required renal replacement therapy (RRT) while 12 (13%) were treated with extracorporeal membrane oxygenation (ECMO), 3/50 (6%) in the SPA and 9/43 (21%) in the IA group (P = 0.06).

Table 1

Main characteristics, biological parameters, anticoagulant treatment and outcome in 93 mechanically ventilated COVID-19 patients.

Parameters	All patients(N = 93)	Standard prophylaxis(N = 50)	Increased anticoagulation(N = 43)	P
Patient characteristics
Male gender, N (%)	64 (69)	36 (72)	28 (65)	0.51
Age (years)	63 [56–71]	62 [54–69]	65 [58–73]	0.14
Body mass index (kg/m2)	29 [25–32]	28 [25–31]	30 [25–34]	0.18
Past hypertension, N (%)	49 (53)	23 (46)	26 (60)	0.21
Diabetes, N (%)	36 (39)	22 (44)	14 (33)	0.29
Ischemic heart disease, N (%)	11 (12)	9 (18)	2 (5)	0.58
SOFA score on admission	6 [3–8]	6 [4–9]	5 [3–8]	0.14
Main biological parameters
PaO2/FiO2 (mmHg)	151 [113–240]	179 [117–258)]	141 [110–188]	0.15
PT (ratio of normal)	1.18 [1.12–1.27]	1.17 [1.11–1.25]	1.19 [1.14–1.31]	0.17
APTT (ratio of normal)	1.23 [1.12–1.50]	1.21 [1.10–1.43]	1.29 [1.20–1.65]	0.04
Plasma fibrinogen (g/L)	7.6 [6.2–8.6]	8.1 [6.7–8.8]	7.2 [6.1–8.1]	0.07
Plasma D-dimer (ng/mL)	3180 [1495–5808]	3500 [2000–7760]	2710 [1465–4135]	0.10
White blood cells (G/L)	10.4 [7.8–14.0]	10.5 [8.0–13.8]	9.8 [7.1–14.3]	0.57
Lymphocytes (G/L)	0.72 [0.42–1.20]	0.74 [0.45–1.14]	1.0 [0.43–1.24]	0.74
Platelets (G/L)	274 [197–367]	271 [200–365]	274 [194–372]	0.99
CRP (mg/L)	223 [132–307]	246 [180–304]	180 [117–307]	0.18
Serum creatinine (µmol/L)	98 [67–154]	94 [70–150]	99 [63–173]	0.87
Serum ALT (IU/L)	33 [23–50]	32 [21–48]	34 [25–54]	0.46
Anti-COVID-19 and supportive treatments
Lopinavir/ritonavir combination, N (%)	12 (13)	12 (24)	0 (0)	0.003
Azithromycin, N (%)	42 (45)	16 (32)	26 (61)	0.01
Hydroxychloroquine, N (%)	25 (27)	14 (28)	11 (26)	0.82
Dexamethasone, N (%)	41 (44)	13 (26)	28 (65)	0.0002
Vasopressor treatment, N (%)	45 (49)	27 (54)	18 (42)	0.21
Renal replacement therapy, N (%)	29 (31)	14 (28)	15 (35)	0.51
ECMO, N (%)	12 (13)	3 (6)	9 (21)	0.06
Anticoagulation regimen
Standard prophylaxis before initial ultrasound, N (%)	50 (54)	50 (100)	0 (0)	<0.0001
Standard prophylactic enoxaparin, N (%)	42 (45)	42 (84)	0 (0)	<0.0001
Standard prophylactic unfractionated heparin, N (%)	8 (9)	8 (16)	0 (0)	0.05
Double-dose prophylactic enoxaparin, N (%)	25 (27)	0 (0)	25 (58)	<0.0001
Therapeutic anticoagulation before initial ultrasound, N (%)	18 (19)	0 (0)	18 (54)	<0.0001
Therapeutic enoxaparin before initial ultrasound, N (%)	6 (6)	0 (0)	6 (14)	0.01
Therapeutic unfractionated heparin before ultrasound, N (%)	12 (13)	0 (0)	12 (28)	<0.0001
Endpoints
Femoral/popliteal DVT, N (%)	15 (16)	13 (26)	2 (5)	0.01
Femoral/popliteal DVT, on enoxaparin prophylaxis, N (%)	13 (14)	11 (22)	1 (2)	0.02
DVT below the popliteal level, N (%)	23 (25)	12 (24)	11 (26)	1.0
Major bleeding, N (%)	18 (19)	7 (14)	11 (26)	0.19
Therapeutic anticoagulation at major bleeding, N (%)	16 (17)	9 (18)	8 (19)	1.0
Death, N (%)	38(44)	18 (34)	20 (56)	0.08

DVT, deep vein thrombosis; SOFA score, Sepsis-related Organ Failure Assessment score; PaO2/FiO2, oxygen arterial partial pressure/fraction of inspired oxygen ratio; PT, prothrombin time; APTT, activated partial thromboplastin time; CRP, C-reactive protein; ALT, alanine-aminotransferase; ECMO, extracorporeal membrane oxygenation.

Prevalence of femoral/popliteal DVT was significantly reduced in the IA in comparison with the SPA group (two (5%) versus 13 (26%), P = 0.01; Fig. 1 ). The two DVT in the IA group and one DVT in the SPA group were associated with femoral central venous catheters. After adjustment for parameters significantly different between groups, anticoagulant treatment was the only factor associated with DVT (P = 0.02). Prevalence of femoral/popliteal DVT was decreased, i.e. 1/25 patients treated with enoxaparin 0.4 mg twice/day (2% of the IA group) versus 11/42 patients treated with enoxaparin 0.4 mg/day (22% of the SPA group), P = 0.02.

Fig. 1

Before-after comparison of proximal deep vein thrombosis (DVT) and major bleeding (MB) prevalence in 93 mechanically ventilated COVID-19 patients.

MB occurred 10 days [8-13] post-intubation in 11 patients (26%) in the IA group versus seven patients (14%) in the SPA group (P = 0.19). MB occurred while on ECMO and/or RRT in 15/18 cases (83%). One patient died of intracranial hemorrhage. In patients treated with 0.4 mg/day enoxaparin in the SPA group, one MB occurred versus none in patients treated with enoxaparin 0.4 mg twice/day in the IA group. In the SPA group, 17/50 (34%) died as compared to 19/43 (44%) in the IA group (P = 0.08), while 7/43 (16%) are still hospitalized.

Our most important finding is that IA is associated with decreased femoral/popliteal DVT prevalence in comparison with SPA in mechanically ventilated COVID-19 patients. The second important finding is that enoxaparin 0.4 mg twice daily regimen seems effective for DVT prophylaxis compared with standard enoxaparin prophylaxis.

To our knowledge, this is the first study comparing SPA with IA strategies and showing a significant reduction in femoral/popliteal DVT using systematic ultrasound screening. Our data suggests that double-dose enoxaparin prophylaxis (40 mg twice daily) may have a favorable risk/benefit ratio, worth exploring in further studies. DVT below the popliteal level were not reduced in the IA group, suggesting that at this level, venous stasis and/or endothelial lesion may play a more important role than hypercoagulation.

Our study strength is that ultrasound was performed in all patients, avoiding biases related to the absence of systematic screening. Limitations include absence of randomization and small sample size precluding assessment of the effect on mortality.

In conclusion, using systematic ultrasound screening, we observed a decrease in femoral/popliteal DVT prevalence while increasing anticoagulation compared to standard prophylaxis in mechanically ventilated COVID-19 patients. The favorable risk/benefit ratio of prophylactic double-dose enoxaparin 40 mg twice-daily regimen is worth exploring in future studies.

Declaration of Competing Interest

The authors declare no competing interests.