Background

Venous thromboembolism (VTE) comprises pulmonary thromboembolism (PTE) and deep vein thrombosis (DVT), with an annual proportion ranging from 0.75 to 2.69 per 1000 people among the populations of Europe and North America [1]. Previous research reported that the age and sex-adjusted proportion of VTE among China’s population increased from 3.2 to 17.5 per 100,000 people from 2007 to 2016 [2]. Under appropriate prophylaxis measures, the proportion of VTE for inpatients has been found to decrease from about 4.9–14.9% to 2.7–5.5% [3, 4]. Patients with critical illness are at high risk of VTE [5, 6]. A previous study showed that the rates of DVT ranged from 13 to 31% in critically ill patients without prophylaxis measures [7]. Moreover, the frequency of VTE in patients in the intensive care unit (ICU) receiving thromboprophylaxis ranged from 5.1 to 15.5% [8]. Many studies have also proposed that VTE events are associated with a poor prognosis in critically ill patients [9, 10].

According to the present guidelines, thromboprophylaxis practice should be done on the basis of a VTE risk evaluation [5, 6]. However, a multinational cross-sectional study reported that 51.8% of hospitalized patients were at risk of VTE, with only one half of those receiving prophylaxis complying with related guidelines [11]. A previous study showed that the overall rate of guidelines-recommended prophylactic method was 10.3% in surgical and medical patients [12]. Meanwhile, the current status of VTE prophylaxis in ICUs is also not optimistic [9, 13]. Our previous study suggested that the awareness rate of VTE prophylaxis among the medical staff of ICUs in North China remains limited, which may lead to a lack of standardized VTE prophylaxis [14]. However, the proportion of major bleeding in critically ill patients under heparin thromboprophylaxis has been found to be about 4–6% [15, 16], which may limit VTE prophylaxis practice in ICUs.

Since August 2014, a thromboprophylaxis protocol has been applied in the respiratory ICU of Beijing Chao-Yang Hospital. This protocol was drawn up based on the relevant guideline and experts’ advice [5, 6] and considers the specialty of the respiratory ICU. This study aimed to explore the effects of VTE prophylaxis on the proportion of VTE, and the risk factors of VTE among critically ill patients in the ICU after receiving the thromboprophylaxis protocol.

Methods

Study design and patients

This study was a single-center, prospective cohort study. Patients admitted to the respiratory ICU from August 1, 2014, to December 31, 2020, were enrolled in this study. Patients aged 18 or older with a length of stay (LOS) in the ICU of more than 48 hours were included in this study. The exclusion criteria were admission to the ICU because of acute PTE and/or proximal DVT event, readmission in 48 hours after transferring out of the ICU, and refusal to participate in the study. This study was reviewed and approved by the Ethics Committee of Beijing Chao-Yang Hospital (2014-Ke-142). Informed consent was obtained from the patients or their legal guardian.

Thromboprophylaxis protocol

The thromboprophylaxis protocol was conducted in a 16-bed respiratory ICU. All critically ill patients in the respiratory ICU were considered as being at high risk of VTE and should receive VTE prophylaxis. Bleeding risk was evaluated first. If the patients had a high bleeding risk or already had active bleeding, mechanical prophylaxis measures were applied. Pharmacological prophylaxis measures were used on patients with a low bleeding risk. If these patients were immobile, such as in deep sedation, on a neuromuscular blocker, or paralyzed, they were stratified as being at extremely high risk of VTE, and pharmacological prophylaxis combined with mechanical measures were conducted. While the patients acquired active bleeding or bleeding risk increased during pharmacological prophylaxis, mechanical measures would be switched instead of pharmacological measures. The risk of bleeding or active bleeding needed dynamic evaluation, and pharmacological prophylaxis had to replace mechanical prophylaxis once the bleeding risk was relieved or active bleeding stopped. Before mechanical prophylaxis was implemented, compression ultrasonography (CUS) of lower extremity had to be conducted. Intermittent pneumatic compression (IPC) was not used while DVT existed; otherwise, IPC combined with graduated compression stocking (GSC) was used. The thromboprophylaxis protocol was maintained until the VTE risk was relieved or there was a new occurrence of VTE events that required therapeutic anticoagulation or thrombolysis (Fig. 1).

Fig. 1
figure 1

Thromboprophylaxis protocol. CUS, compression ultrasonography; GCS, graduated compression stocking; IPC, intermittent pneumatic compression; VTE, venous thromboembolism

Full size image

For the aspect of pharmacological prophylaxis, low molecular weight heparin (LMWH) with a prophylactic dose was the first choice in patients without contraindications. Enoxaparine with the dose of 40 mg subcutaneous injection per day was chosen for pharmacological prophylaxis. The dosage would be adjusted if the patients with extreme weight. Unfractionated heparin (UFH) was used as 5000 IU subcutaneous injection twice per day or continuous intravenous infusion with range of APTT about 1.2 times baseline was the alternative anticoagulant in patients with a non-bleeding contraindication of LMWH. If patients had a history of heparin-induced thrombocytopenia (HIT) or antithrombin deficiency, a prophylactic dose of a non-heparin anticoagulant was used, such as fondaparinux with 2.5 mg subcutaneous injection per day. If the patients had specific medical histories such as atrial fibrillation, chronic thromboembolic pulmonary hypertension, or post-cardiac surgery, the dosage of anticoagulation had to be according to the disease-specific treatment need. The patients underwent extracorporeal support including extracorporeal membrane oxygenation (ECMO) or continuous renal replacement therapy (CRRT), and the anticoagulation management had to be in line with the relevant protocol.

If patients with an active gastroduodenal ulcer, prior bleeding history in the 3 months before admission, low platelet count (less than 5 × 109/L), hepatic failure (international normalized ratio higher than 1.5), and activated partial thromboplastin time (APTT) increased (10 s increased) without an anticoagulation agent, they were assessed as being at high risk of hemorrhage [16, 17].

The CUS examination was conducted as a screening of DVT in the first 24 hours after admission to ICU. During ICU hospitalization, if the patients presented with suspicious clinical manifestations of PTE or DVT, diagnosis and treatment process had to be according to the related guidelines [18]. If patients did not show any suspicious clinical symptoms, CUS was re-examined to avoid missing asymptomatic DVT before discharge from the ICU or death. DVT events referred to newly formed sites of lower extremity DVT.

Outcomes

The primary outcome was newly developed VTE events during the ICU stay. Secondary outcomes comprised bleeding events, thrombocytopenia, all-cause mortality in the ICU, and length of ICU stay. Bleeding events included gastrointestinal bleeding, urinary tract bleeding, oral or nasal bleeding, lower respiratory tract bleeding, retroperitoneal bleeding, skin bleeding, intracranial bleeding, surgical incision bleeding, and vaginal bleeding. Major bleeding events were defined as hemoglobin decline ≥2 g/L, hemorrhage treated by blood transfusion of more than 2 units of red blood cells, retroperitoneal hemorrhage, intracranial hemorrhage, hemorrhagic shock, and fatal hemorrhage [19]. Thrombocytopenia was defined as a 30–50% reduction in the baseline platelet level [20]. Thrombocytopenia was recorded at any time during thromboprophylaxis and was analyzed to ascertain the causes by physicians [21,22,23]. If any anticoagulant-related thrombocytopenia occurred, the suspicious drug was ceased.

Clinical data collection

Demographic and clinical data of the patients were entered into an electronic case report form and included the following: demographic characteristics (age and sex), diagnosis, comorbidities, complications, laboratory tests (e.g., routine blood test, coagulation function, liver function, renal function), and organ support. The Caprini score [24] and Padua score [5] were recorded during admission. The VTE prophylaxis measures, proportion of VTE and bleeding events, ICU mortality, and length of ICU stay were also documented.

Statistics analysis

Statistical analysis was performed with IBM SPSS 26.0. Categorical variables were described as frequency (percentage), and differences between groups were tested by the Chi-square test and Fisher’s exact test. Continuous variables were described by the median (interquartile range [IQR]) because of the non-normality distribution. Differences between groups were tested by the Mann–Whitney U test or Kruskal–Wallis H test. Univariate and multivariate logistics regression analyses were conducted to explore the risk factors of VTE events despite being under this protocol. The multivariable regression model was adjusted for the following characteristics: age, D-dimer, blood component infusion, shock, LOS in ICU, CRRT, artificial airway, and cessation of thromboprophylaxis. Variables with P < 0.1 in univariate analysis were included in multivariable logistic regression analysis. P < 0.05 was considered statistically significant.

Results

Patients’ characteristics and VTE prophylaxis

From August 1, 2014, to December 31, 2020, there were 1057 patients admitted to the respiratory ICU. Overall, a total of 943 cases were ultimately screened in the study (Fig. 2). Of these, 59 (6.3%) cases did not undergo any VTE prophylaxis (Supplementary Table 1). Finally, 884 patients received the thromboprophylaxis protocol, with the rate of VTE prophylaxis being 93.7%. Furthermore, 10.5% of patients received mechanical prophylaxis, 43.8% underwent pharmacological prophylaxis, and 45.7% received pharmacological prophylaxis combined with mechanical prophylaxis (Table 1).

Fig. 2
figure 2

Flow chart. DVT, deep vein thrombosis; LOS, length of stay; IDDVT, isolated distal deep vein thrombosis; PTE, pulmonary thromboembolism; RICU, respiratory intensive care unit; VTE, venous thromboembolism

Full size image
Table 1 Characteristics of patients in different prophylaxis groups
Full size table

For patients who underwent VTE prophylaxis, the median age was 61 (IQR 48,70) years, the median BMI was 23.7 (IQR 20.8, 26.7) kg/m2, and 66.9% patients were male. The pharmacological combined with mechanical prophylaxis group had a higher proportion of acute respiratory distress syndrome (ARDS) and pneumonia than the other groups (P < 0.05). The number of patients with acute exacerbation of chronic obstructive pulmonary disease (AECOPD) and invasive pulmonary aspergillosis (IPA) in the pharmacological prophylaxis group was more than that in the other two groups (P < 0.05).

Patients in the mechanical prophylaxis group had lower hemoglobin, platelet, and fibrinogen levels and a higher urea nitrogen level than patients in the other two groups (P < 0.05). Patients in the pharmacological combined mechanical prophylaxis group had a shorter activated partial thromboplastin time and a higher C-reactive protein level than those in the pharmacological prophylaxis group patients (P < 0.05). The d-dimer level of the pharmacological prophylaxis group was significantly higher than that in the other two groups (P < 0.001) (Table 2). The number of patients with an intravascular tube and underwent invasive mechanical ventilation was greater in the pharmacological combined mechanical prophylaxis compared to other measures (P < 0.001) (Table 3).

Table 2 Laboratory tests of patients in different prophylaxis measures
Full size table
Table 3 Treatments for different prophylaxis groups during thromboprophylaxis
Full size table

VTE event of different prophylaxis measures

The total proportion of VTE events was 14.3% (126/884) for patients who underwent VTE prophylaxis, of which 0.1% (1/884) had PTE, 2.0% (18/884) had proximal DVT, and 12.1% (107/884) had isolated distal deep vein thrombosis (IDDVT). The proportions of proximal DVT or IDDVT among different prophylaxis groups were not significantly different (Table 4).

Table 4 Outcomes for different prophylaxis
Full size table

Bleeding and thrombocytopenia event of different prophylaxis measures

About 16.2% (143/884) patients had bleeding events under the VTE prophylaxis protocol during the ICU stay, and 9.4% (83/884) of them were major bleeding events (Table 4). There were no significant differences in the proportion of bleeding events in different prophylaxis groups (P = 0.683). During thromboprophylaxis, five patients had anticoagulant-related thrombocytopenia, with no significant difference among different prophylaxis groups.

ICU mortality and length of ICU stay

Mortality for patients who underwent VTE prophylaxis protocol was 29.0% (256/884), but there was no significant difference between different prophylaxis groups (P = 0.167) (Table 4). For patients with VTE, the mortality rate was 34.1% (43/126). There was no difference in mortality between patients whether or not they experienced VTE events during their ICU stay (P = 0.169). The LOS in the ICU of patients in the pharmacological combined mechanical prophylaxis group was 14 (9, 25) days, which is significantly longer than that in the other two groups (P < 0.001). The duration of the ICU stay of patients with VTE events was 20 (12, 35) days, which is significantly longer than in patients without VTE events of 14 (8, 26) days, P < 0.001.

Risk factors for VTE under the prophylaxis protocol

Multivariate logistic regression revealed hepatic failure due to cirrhosis (OR 5.789, 95% CI [1.402, 23.894], P = 0.015), acute asthma exacerbation (OR 39.999, 95% CI [4.704, 340.083], P = 0.001), and ECMO (OR 22.237, 95% CI [4.824, 102.502], P < 0.001) were independent risk factors of proximal DVT of patients in the ICU under the VTE prophylaxis protocol (Fig. 3 and Supplementary Table 2). For the aspect of IDDVT, artificial airway (OR 2.886, 95%CI [1.551, 5.372], P = 0.001) and duration of mechanical ventilation (OR 1.020, 95%CI [1.010, 1.029], P < 0.001) were the independent risk factors of patients under VTE prophylaxis (Supplementary Table 3).

Fig. 3
figure 3

Forest plot of multivariate regression of proximal DVT. AST, aspartate aminotransferase; CRRT, continued renal replacement therapy; ECMO, extracorporeal membrane oxygenation

Full size image

Discussion

The present study was the largest cohort study in mainland China focused on the VTE event proportion under a thromboprophylaxis protocol for critically ill patients in ICU. The thromboprophylaxis rate in this study was extremely high. However, there was still a certain occurrence of VTE events under this thromboprophylaxis protocol, albeit with a relatively lower proportion rate of PTE and proximal DVT. In this study, we found that the VTE events that happened were not associated with the increased risk of mortality among critically ill patients with thromboprophylaxis. Hepatic failure due to cirrhosis, acute exacerbation of asthma, and ECMO were the independent risk factors of proximal DVT despite being under the thromboprophylaxis protocol. Nevertheless, about 16.2% of the patients had bleeding events under this prophylaxis protocol, which should be paid more attention to during implementation.

Critically ill patients generally have more than one VTE risk factor [25]; therefore, thromboprophylaxis is recommended for such patients in many sets of guidelines [5, 6, 26]. The PROF-ETEV study from Spain found that about 41% of critically ill patients were receiving an inappropriate prophylaxis [13]. Research from Australia showed that early thromboprophylaxis was used in 74% patients within 24 h of ICU admission [9]. An observation study has reported that with the increase of the medical staff’s awareness of VTE prophylaxis in China, the VTE prophylaxis rate in ICUs was about 90.1% [27]. In the present study, the thromboprophylaxis rate in the ICU was 93.7%, which is relatively higher than that mentioned in previous literature.

A multi-center study showed that proximal DVT occurred in 5.1% of critically ill patients receiving LMWH [28]. Hamada et al. found that the prevalence of VTE was still 30.7% despite the well-driven thromboprophylaxis protocol in critically ill trauma patients [29]. A pooled analysis of two prospective cohort studies discovered that 2.2% of the critically ill patients receiving contemporary thrombosis prophylaxis developed PTE with or without DVT [10]. In this study, all the critically ill patients received thromboprophylaxis protocol once admitted to the ICU. The specific thromboprophylaxis measure selection should be done according to the bleeding risk and whether there exists an extremely high risk of VTE. We found that the proportion of proximal DVT and PTE was lower than that in the existing literature reported under this protocol. Therefore, this thromboprophylaxis protocol seemed to be more effective, which may indicate its good application prospects.

From the thromboprophylaxis measure distribution characteristics reported in this study, it could be found that nearly 90% of critically ill patients in a non-surgical ICU should receive at least pharmacological prophylaxis. There was no difference in the proportion of VTE events of patients between different thromboprophylaxis measures in this study. Thromboprophylaxis by mechanical alone is recommended for critical care patients at high risk of bleeding or active bleeding with contraindications to prophylactic anticoagulant agents because of the uncertain benefit of mechanical prophylaxis measures [5]. However, there has still been a lack of research on the thromboprophylaxis effect between pharmacological and mechanical prophylaxis measures. Although there were fewer patients in the mechanical prophylaxis group than in the other groups, this may hint at the equal effect of thromboprophylaxis according to the bleeding risk stratification in this protocol. It also indicates the rationality of this protocol.

Except for existing VTE risk assessment models, it is important to recognize the risk factor of VTE despite being under specific thromboprophylaxis. In this study, cirrhosis, asthma, and ECMO were the independent risk factors of proximal DVT under the thromboprophylaxis protocol. Some literature suggests that patients with cirrhosis have an increased risk of VTE [30]. The possible mechanisms are the reduction of anticoagulant factors, hyperactivation of thrombin, procoagulant activity caused by structural changes of fibrin, and platelet hyperreactivity in patients with cirrhosis. Impaired fibrinolysis might also be a mechanism for the increased risk of VTE in asthma patients [31], and this risk was higher in younger and more severe asthmatic patients [32]. With the increasing application of ECMO in clinical settings, ECMO-related VTE event has been found to range from 18.1 to 74% [33, 34]. The mechanism of thrombosis in ECMO may be not only with the local endothelial injury but also the difficulty in management of anticoagulation of ECMO.

In the present study, thromboprophylaxis protocol could effectively decrease the occurrence of proximal DVT and PTE, but 12.1% of the patients still acquired asymptomatic IDDVT. Currently, there is still controversy regarding the clinical benefit of treatment and long-term prognosis of IDDVT [35, 36], because there are very little data available on critically ill patients. We found that the presence of an artificial airway and the duration of mechanical ventilation were independent risk factors of IDDVT. This might be because early mobilization was limited by long time mechanical ventilation, which caused the muscle group of distal lower limb contraction weaken with slow venous reflux. Decreasing the proportion of IDDVT in ICU patients might be another important research topic in the future.

In this study, 9.4% of the patients received anticoagulation agents for thromboprophylaxis and suffered major bleeding events. The proportion of major bleeding events did not differ between different thromboprophylaxis measures. However, the prevalence of major bleeding in this study was higher than that in other LMWH-related studies [16]. A systematic review and meta-analysis showed that major bleeding did not appear to be significantly influenced by heparin thromboprophylaxis in the ICU setting [15]. Another high-risk factor for bleeding seems to be patients on ECMO [37]. Indeed, patients who underwent ECMO are with high risk of bleeding and thrombosis at same time. Meanwhile, when patients on ECMO were excluded from analysis in this study, the proportion of major bleeding decreased to 6.2%, which is in line with existing reports.

There were some limitations in this study. First, this was a single-center cohort study, which might induce an unavoidable selection bias. Second, independent risk factors of major bleeding under this thromboprophylaxis protocol were not analyzed, which should be the next important work of this cohort study. Third, during the present study, thromboprophylaxis measures transitioned in different arms according to the risk of bleeding and VTE changes during ICU stay. It had been difficulty in grouping patients exactly, which might affect the power of the result. We tried to minimize these interfere by grouped the patients with the longest duration of the prophylaxis measures before end-point events. Lastly, this study was conducted in a respiratory ICU, and the result could only be spread among non-surgical critically ill patients. Although this thromboprophylaxis protocol was carried out based on the assessment of bleeding and VTE risks, its safety and effect should be further explored, not only in an enlarged sample size study but also in different ICUs and regions.

Conclusion

The thromboprophylaxis protocol for critically ill patients drawn up based on guidelines was practicable in the ICU, which with the potential to help reduce the proximal VTE and PTE event proportions. However, this protocol may have an unsatisfactory effect in some special patients. Therefore, it is important to recognize the risk factors of VTE events happening under the thromboprophylaxis protocol. Early intervention or strengthening of prophylaxis measures may help to reduce the risk of VTE in such patients. Meanwhile, anticoagulation agents related to major bleeding should be monitored while implementing this protocol. Further study should focus on these factors to perfect the thromboprophylaxis protocol in the future.