Prevalence of DVT in patients with SCAP undergoing thromboprophylaxis

In recent years, there have been numerous studies on COVID-19 related VTE or DVT, but few papers on VTE or DVT caused by non-COVID-19 related SCAP, especially in patients who had already received VTE prophylaxis. In this single-center study, we found a high incidence of DVT in SCAP patients undergoing thromboprophylaxis recommended by guideline [20], with no significant difference in the incidence of both COVID-19 and non-COVID-19 induced SCAP. The occurrence of DVT is independently associated with personal history of VTE, longer bedridden time, higher D-dimer levels, higher LDH levels, and IMV.

In 2002, Greets et al. reported that the rates of objectively confirmed DVT in 4 prospective studies ranged from 13 to 31% and suggested a potential role of thromboprophylaxis in patients requiring critical care [30]. Follow-up studies and guidelines had also confirmed the value of VTE prophylaxis [20, 31], but some recent studies have also shown that, despite receiving guideline-recommended thromboprophylaxis [20], the incidence of DVT is still as high as 14% to 37.2% in critically ill patients [16,17,18,19].

In our study, even with VTE prophylaxis, the incidence of DVT in patients with SCAP was as high as 35.6%. Several reasons probably account for the high prevalence of DVT in our study. First, most of the above studies generally focused on critically ill patients. SCAP is a more serious type of critical illness that shows an overwhelming systemic inflammatory process accompanied by vascular endothelial injury and hypercoagulability, and may have a higher risk of DVT [5, 12,13,14,15, 17]. Second, immobility in these patients leads to stasis of blood flow, further intensifying their predisposition to DVT [11]. Third, while some studies define DVT as either proximal DVT or symptomatic DVT [6, 32], yet, in our research, we conducted lower extremity venous compression ultrasonography on all patients regardless of whether they exhibited symptoms of DVT, which might increase the detection rate of DVT.

Risk factors for DVT in patients with SCAP undergoing thromboprophylaxis

Our study shows that the history of VTE, longer bedridden times, higher D-dimer levels, higher LDH levels, and IMV were independent risk factors for DVT. Using a ROC analysis, the combination of VTE history, bedridden times, D-dimer levels, LDH levels and IMV yielded a sensitivity of 80.6% and a specificity of 81.4% for scanning for DVT in these hospitalized patients.

Aligning with prior studies and guidelines [33,34,35], our survey indicates that a history of VTE is an independent risk factor for the development of DVT. Despite the implementation of prudent preventive measures, the existence of a prior VTE history significantly contributes to the risk of DVT occurrence (with an adjusted OR as high as 13, which is a striking statistic). A history of VTE is a potentially strong risk factor for recurrent VTE events [11]. As early as the beginning of this century, Baglin and Hansson et al. highlighted that patients with a history of VTE had a 2- and 5-year recurrence rate of VTE higher than 10% [36, 37]. Subsequent research has also affirmed that a personal history of VTE stands as an independent risk factor for VTE recurrence, and this factor has been recognized as a non-modifiable risk factor, been integrated into various VTE risk assessment tools and authoritative guidelines [11, 25, 26, 38]. This may be attributable to the potential for underlying genetic and acquired thrombophilia in such patients, conferring a heightened likelihood of thrombus recurrence. It is regrettable that in our study, the cohort of patients with a personal history of VTE was limited to only 13 individuals, and we did not conduct a follow-up on their thrombophilia, particularly their genetic predisposition to VTE. This represents an area that warrants further investigation in future research.

Prolonged bed rest was also identified as a risk factor for the development of DVT in this study. Even with the implementation of reasonable preventive measures, patients who have been bedridden for more than 3 days face a 17-fold increase in the risk of DVT. As one of the significant contributing factors to the occurrence of VTE [39, 40], prolonged bed rest or immobilization has been incorporated into various VTE risk assessment models and authoritative guidelines [11, 25, 26, 38]. This may be because bed rest leads to a reduction in physical activity, which in turn slows the return of blood from the lower extremities, causing stasis and thereby increasing the risk of thrombus formation. Furthermore, an extended period of bed rest often indicates a more severe condition in these patients, potentially accompanied by a more serious inflammatory response and a hypercoagulable state, which in turn raises the risk of developing DVT.

Within our research, it was discerned that elevated D-dimer levels at admission is indicative of an increased risk for DVT among patients afflicted with SCAP who were administered prophylaxis against VTE. D-dimer is a fibrin degradation product encompassing multiple cross-linked D domains and/or E domains present in the original fibrinogen molecule, whose generation is only theoretically possible when hemostasis and fibrinolysis pathways are concomitantly activated. Measurement of plasma D-dimer level has now become a basic element in the diagnosis or exclusion and prognostication of VTE when incorporated into validated clinical algorithms. In a comprehensive study by Tsaplin et al., a cohort of 168 patients with COVID-19 was evaluated to discern the efficacy of the original versus the revised Caprini scoring system in assessing the risk of VTE. The latter, which includes elevated D-dimer levels, demonstrated superior accuracy in predicting VTE risk among the studied population [41]. It is a well-documented phenomenon that patients afflicted with CAP, irrespective of whether the etiology is viral or bacterial, frequently present with increased D-dimer levels and an augmented risk of DVT [16, 42, 43], a concern that is particularly pronounced in those SCAP. Among patients exhibiting elevated D-dimer levels at admission, vigilant surveillance is paramount, notwithstanding the implementation of VTE preventative measures. This proactive approach is essential for the timely identification of DVT and the subsequent initiation of appropriate therapeutic measures.

We had concurrently identified an association between elevated serum LDH levels and the occurrence of DVT. LDH, a pivotal enzyme ubiquitously present within nearly every tissue of the human body, including red blood cells, heart, lungs, liver, muscles, and kidneys, plays an indispensable role in cellular metabolism by catalyzing the interconversion of pyruvate to lactate and vice versa [44]. Upon the onset of tissue injury or infection, LDH is expressed by the affected tissues or cells, leading to an increase in serum LDH levels. Consequently, elevated LDH levels are observed in a myriad of conditions, including infections, and are often correlated with the severity and prognosis of the disease [45, 46]. In concordance with previous findings in patients with COVID-19 [47,48,49], we had observed a higher expression of serum LDH in SCAP patients with concomitant DVT. The likely rationale for this phenomenon is the occurrence of DVT, which, due to impeded blood circulation, results in localized tissue ischemia, hypoxia, edema, degeneration, and potentially necrosis. This cascade of events leads to the release of intracellular LDH into the bloodstream, thereby manifesting as elevated serum LDH levels [50].

The utilization of IMV emerged as an additional independent risk factor associated with the occurrence of DVT in our study. For critically ill patients, the risk of hospital-acquired VTE escalates during the course of IMV treatment, with an increasing duration of IMV conferring a progressively higher risk for VTE [32, 51, 52]. Due to the typical presence of severe respiratory failure, the necessity for IMV in the context of SCAP is associated with an increased predisposition towards VTE. As supported by literature reports and our statistical analysis, prolonged immobilization inherent to IMV potentiates the risk for VTE. The hypercoagulable state induced by critical illness, exacerbated by the inflammatory cascade of severe pneumonia, also augments the likelihood of thrombotic events. Nonetheless, within the parameters of our investigation, a discernible statistical correlation was not detected between the duration of IMV and the incidence of DVT. This absence of correlation could be ascribed to the relatively small sample size of our patient cohort.

ROC and nomogram for screening for DVT in patients with SCAP undergoing thromboprophylaxis

The susceptibility to VTE intensifies with an increasing number of precipitating factors [11]. In the context of SCAP who have received VTE prophylaxis recommended by guidelines, the newly formulated predictive model demonstrates superior efficacy in forecasting the incidence of DVT. Existing models, such as the Caprini prediction score and Padua prediction score, which are prevalent in clinical practice, are designed for the general inpatient population that has not yet undergone thromboprophylaxis and do not account for specific medical conditions. The novel model, specifically crafted for the cohort of SCAP patients who have already been subjected to preventive measures for VTE, and which integrates characteristic indicators of the patients' condition, such as the requirement for IMV, provides a more precise prediction of DVT occurrence within this particular group. This enhanced predictive acumen could potentially facilitate the early identification of individuals at heightened risk, thereby allowing for the refinement and optimization of their preventive measures.

Prognosis of DVT in patients with SCAP undergoing thromboprophylaxis

Similar to prior investigations [16, 19, 43], our findings suggested a connection between DVT and a bleaker prognosis for patients. Notably, the 30-day and in-hospital mortality rates are markedly higher among patients with concurrent DVT compared to those without DVT. Univariate analysis revealed that patients with DVT exhibit elevated neutrophil counts and D-dimer levels compared to patients without DVT, indicating a potential escalation in inflammatory response and subsequent exacerbation of coagulation and fibrinolysis dysfunction [12, 14]. Additionally, individuals in the DVT cohort experience prolonged periods of immobility, lower oxygenation indices, higher rates of IMV therapy, and elevated severity scores such as APACHE II and SOFA, suggesting that these patients may represent a more severe subset of SCAP. In addition, there is a 50% chance for patients with untreated proximal DVT to develop symptomatic PE [53] which might aggravate the hypoxemia of SCAP patients and then result in lower actuarial survival rates. If there is any clinical suspicion of PE, a CTPA should be considered and obtained. Unfortunately, due to the critical condition of SCAP, CTPA examination was limited. CTPA examination was performed only for 11 patients with SCAP who were highly suspected of PE and 9 patients consequently diagnosed with PE. Using the figures given above, we may underestimate the incidence of PE. The potential presence of PE associated with DVT may be also associated with poor survival in patients with DVT. While this scenario may not be surprising given that our population consists of critically ill patients with poor prognosis, the high incidence of DVT and its associated worsened clinical outcomes, even in those who have received VTE prophylaxis, underscores the importance of screening for DVT-related risk factors in such patients. It is imperative to explore specific and optimized VTE prevention strategies to improve the prognosis of these individuals.

In this study, we identified high-risk patients for DVT among SCAP patients using a risk assessment model and explored the value of ultrasound screening. Although routine lower limb venous duplex ultrasound screening may face challenges in resource-limited settings, selecting high-risk patients for ultrasound screening based on a risk assessment model can improve diagnostic efficiency without adding excessive resource burdens. Moreover, while routine ultrasound screening may not dramatically alter clinical practice, identifying high-risk patients and implementing targeted preventive measures can significantly reduce the incidence and mortality of DVT. For example, the 30-day mortality and in-hospital mortality were significantly higher in DVT group than those in non-DVT group in our study, highlighting the importance of early identification and intervention. By optimizing resource allocation, ultrasound screening can achieve better clinical outcomes in resource-limited settings.

Despite these findings, our study is not without its inherent limitations. First, the retrospective nature of our research precluded the tracking of genetic predispositions in patients afflicted with DVT. Second, due to the same retrospective design, an analysis of the initial causative factors in patients with a prior history of VTE was not conducted. Third, given the critical condition of SCAP patients, the application of CTPA was constrained, resulting in only 11 cases within our study cohort undergone CTPA for suspected PE, which may have led to an underestimation of PE prevalence. Fourth, while the incidence of DVT was marginally lower in patients treated with a combination of LMWH and mechanical prophylaxis compared to those with LMWH alone (30.0% vs 38.6%), the difference was not statistically significant (P = 0.172). A larger sample size might yield more compelling results. Additionally, the small sample size (309 patients) may have led to wide CIs for some variables, increasing the uncertainty of the results. Although we attempted to account for potential confounders in the multivariate analysis, we cannot completely rule out the influence of other unknown factors on the results. Finally, although the predictive model performed well in our dataset (AUC = 0.856), it has not yet been validated in an independent cohort, and its generalizability to different populations remains to be further assessed. Currently, we are collecting additional data and plan to validate the predictive model in an independent cohort in future studies. This will help further optimize the model and assess its applicability in different clinical settings.