The Prognostic Value of D-Dimer Levels for Injury Outcomes in Trauma Patients: A Systematic Review and Meta-Analysis

Hooman Rezaei (PubMed)(Google Scholar)

Maryam Hosseini (PubMed)(Google Scholar)

Introduction

Trauma is a leading cause of global mortality and disability, representing the sixth most common underlying cause of mortality and the fifth leading cause of moderate and severe disability worldwide. It results in the highest rates of moderate and severe disability and mortality among young people [1, 2]. Although the Glasgow Coma Score (GCS) [3, 4] and repeated computed tomography (CT) scans are conventional approaches for evaluating trauma injuries [5-7], reliable prognostic markers are required to assist in the immediate diagnosis of complications and predict patient recovery [8].

Acquired coagulopathy disorders are common consequences of tissue injury, hypoperfusion, hypothermia, and acidosis in trauma patients. These disorders result from an imbalance between anticoagulant and procoagulant factors, platelet dysfunction, and, most importantly, fibrinolysis during the acute phase of trauma. They are associated with a wide range of life-threatening outcomes, from progressive hemorrhagic injury (PHI) to mortality [9-11]. Hyperfibrinolysis is a critical aspect of trauma-induced coagulopathy, characterized by drastically elevated levels of fibrinogen/fibrin degradation products (FDPs), particularly D-dimer. This elevation is integrally linked to poor prognosis and severe outcomes [12, 13]. Previous studies demonstrated that alterations in D-dimer concentration could disrupt the fibrinolytic system by unbalancing coagulation factors, which could ultimately increase the need for massive transfusion due to extreme bleeding [14, 15]. Several investigations evaluated the prognostic role of D-dimer levels in different types of traumatic injury, including traumatic brain injury (TBI), but have failed to reach a decisive conclusion due to discrepant findings [16, 17]. For instance, a prospective study (n=205) found that a D-dimer>1,793 ng/mL at admission predicted mortality (OR=5.87) [18]. Another meta-analysis highlighted D-dimer’s superiority over INR/PT for predicting disseminated intravascular coagulation (DIC) and hemorrhagic progression in TBI [19]. Moreover, a systematic review and meta-analysis by Zhang et al., confirmed the prognostic role of D-dimer levels upon admission in patients with TBI, revealing a direct association between higher D-dimer levels upon admission and the risk of PHI [10]. While the study by Zhang et al., focused exclusively on TBI patients [10], the present study was designed to systematically review and quantify the association between D-dimer levels and injury outcomes in a broader population of trauma patients by conducting a meta-analysis.

Materials and Methods

This study was conducted in accordance with the guidelines of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement.

Search Strategy and Selection Studies

Relevant publications were identified through a comprehensive and systematic search of PubMed, MEDLINE/PubMed, and Web of Science databases. The search strategy utilized keywords such as “D-dimer levels”, “fibrin fragment D”, “traumatic injury”, “trauma”, “outcome”, “mortality”, “deep vein thrombosis”, and “DVT”. The Boolean operators (“AND” and “OR”) were used to combine these terms. The initial screening of records was based on titles and abstracts. The full texts of potentially eligible studies were then independently assessed by two reviewers. Any disagreements were resolved through discussion until a consensus was reached. The search was restricted to English-language articles published between January 2011 and January 2023.

Eligibility Criteria

Studies meeting the following criteria were included in the meta-analysis:

1) Study types: Case-control, cross-sectional, and cohort studies.

2) Population: Trauma patients.

3) Exposure/Outcome: Literature examining the association between D-dimer levels and clinical outcomes, with outcome rates (number or percentage of trauma patients with the outcome) clearly stated.

4) Comparison: Studies reporting D-dimer levels in trauma patients for at least two distinct comparison groups (e.g., poor vs. good outcome, survivors vs. non-survivors).

5) Original studies providing the number of participants in each group, the timing of D-dimer evaluation, and the mean and standard deviation of D-dimer levels for each group.

Studies with incomplete or irrelevant data were excluded from this meta-analysis.

Quality Assessment

The risk of bias was assessed independently by two reviewers using the Newcastle-Ottawa Scale (NOS). The NOS evaluates three domains: subject selection (0-4 points), comparability of subjects (0-2 points), and exposure (for case-control studies) or outcome (for cohort studies) (0-3 points) [20]. Total scores range from 0 to 9, with studies scoring ≥6 being considered high-quality. Any disagreements between the two reviewers were resolved through discussion with a third reviewer to reach a consensus (Table 1). Studies scoring below 6 were considered low-quality and were excluded from the analysis.

Statistical Analysis

The association between D-dimer levels and injury outcomes in trauma patients was examined by synthesizing data from the included studies. The means, standard deviations, and sample sizes from each study were used to calculate the standardized mean difference (SMD). Notably, the control groups in these studies typically consisted of trauma patients with good functional outcomes. Consequently, the D-dimer levels in these control groups were expected to be elevated compared to healthy reference values (<250 ng/mL) [21], which contributed to heterogeneity across the studies. Heterogeneity was assessed using the I² statistic and the Chi-squared (Q) test. The I² values were interpreted as follows: <0.25% low, 0.25-0.75% moderate, and >0.75 high heterogeneity. A random-effects model was employed for the meta-analysis due to the presence of significant heterogeneity. In cases where the I² statistic was 0%, indicating homogeneity, a fixed-effects model was applied. Publication bias was assessed using a funnel plot and Egger’s test. All statistical analyses were conducted using STATA version 17, with a significance level set at p<0.05.

Results

The initial database search identified 84 articles. After removing 56 records due to irrelevance or duplication, 28 studies underwent title and abstract screening. This led to the exclusion of 11 studies. The full texts of the remaining 17 articles were assessed for eligibility, and 5 were excluded due to irrelevant or incomplete data. Consequently, 12 studies were included in the final meta-analysis. The study selection process is detailed in the PRISMA flow diagram (Figure 1).

Fig. 1. The flow diagram shows the study selection strategies according to the PRISMA guidelines.

Study Characteristics and Quality Assessment

The characteristics of 17 studies included in the systematic review are summarized in Table 1. All studies were published between 2011 and 2023 and were conducted in six different countries, primarily in China and Japan. Of the 17 studies, 10 were retrospective [16, 22-30] and 6 were prospective [17, 21, 31-34]. The sample sizes ranged from 50 to 2,570 patients. The pooled study population consisted of 5,042 (68%) men and 2,371(32%) women. The outcomes of interest—including mortality, deep vein thrombosis (DVT), hemorrhage, and DIC—occurred in 981 (13.23%) patients, while 6,432 patients did not experience these outcomes.

Regarding patient populations, nine studies focused on traumatic brain injury (TBI), six on multiple trauma, and five on trauma to the extremities, pelvis, face, thorax, or abdomen. Blood samples for D-dimer measurement were obtained on the first day of admission for all patients.

In the methodological quality assessment, ten studies received a score of ≥7 (mean score=6.76) on the NOS and were considered high-quality. These studies met key criteria, including proper case selection, appropriate control groups, valid exposure/outcome assessment, control of confounders (such as age and trauma severity), and adequate follow-up, indicating a low risk of bias.

Comparison of D-dimer between Patients with and without Poor outcomes

A meta-analysis of 10 studies, using the SMD as the effect size, revealed significant heterogeneity (I²=85.6%, p<0.001). Consequently, a random-effects model was employed. The pooled analysis demonstrated a statistically significant difference in the mean D-dimer levels between patients with poor outcomes (16,570.33 ng/mL) and those without poor outcomes and the control group (8,226.23 ng/mL) (p=0.0003). The SMD between the two groups was calculated at 0.51 (95% CI: 0.24 to 0.79, Figure 2).

Fig. 2. A forest plot shows the standardized mean difference of D-dimer levels between trauma patients with poor outcomes (case) and those without poor outcomes (control).

Comparison of D-dimer between Survivors and Non-Survivors

In the meta-analysis of 5 studies assessing mortality among trauma patients, the SMD also indicated significant heterogeneity (I²=88.46%, p<0.001). A random-effects model was therefore applied. The analysis revealed a statistically significant difference in mean D-dimer levels between non-survivors (16,266 ng/mL) and survivors (7,729.01 ng/mL, p=0.03). The SMD was 0.42 with a confidence interval of 0.04 to 0.79 (Figure 3a).

Fig. 3. A forest plot displays the standardized mean difference of D-dimer levels between trauma patients: (a) deceased (case) and survived (control), (b) with deep vein thrombosis (DVT) (case) and without DVT (control).

Comparison of D-dimer between DVT Patients and Non-DVT Patients

The meta-analysis of 3 studies on DVT demonstrated heterogeneity (I²=75.58%, p=0.01, Q=8.93). A random-effects model was used and indicated a statistically significant difference in mean D-dimer levels between patients with DVT (18,656.12 ng/mL) and those without DVT (12,909.52 ng/mL) (p=0.0008). The SMD was 0.79 (95% CI: 0.32 to 1.25, Figure 3b).

Comparison of D-dimer among Patients Based on Their Type of Trauma

A meta-analysis of five studies on TBI patients revealed significant heterogeneity (I²=92.08%, p<0.001). The results of the random-effects model showed a statistically significant difference in mean D-dimer levels between the TBI patients
(14,151.12 ng /mL) and the control group (70,606.04 ng/mL, p=0.012). The SMD between these groups was 0.53, within a confidence interval of 0.11 to 0.94 (Figure 4a).

Fig. 4. A forest plot presents the standardized mean difference of D-dimer levels among (a) TBI patients, (b) multiple trauma patients, and (c) lower extremity trauma individuals.

The analysis of two studies on multiple traumas demonstrated homogeneity (I²=0%, p=0.97), leading to the use of a fixed-effects model. A statistically significant difference in mean D-dimer levels was found between patients with multiple trauma (58,018.24 ng /mL) and the control group (35,466.23 ng /mL, p<0.001). The SMD was 0.87, with a confidence interval of 0.52 to 1.21 (Figure 4b).

The meta-analysis of three studies on lower extremity injuries demonstrated heterogeneity (I²=75.58%, p=0.01); consequently, a random-effects model was used. The analysis revealed a statistically significant difference in mean D-dimer levels between patients with lower extremity injuries (18,656.12 ng/mL) and controls (12,909.52 ng/mL, p=0.0008). The SMD was 0.79 (95% CI: 0.32 to 1.25, Figure 4c).

Publication bias was assessed using Begg’s and Egger’s tests across the included studies. The results indicated no significant publication bias for the analysis of the prognostic role of D-dimer levels (Egger’s test p=0.87, Figure 5).

Fig. 5. Funnel plot assessing the publication bias

Discussion

This systematic review and meta-analysis confirmed that elevated D-dimer levels upon hospital admission are a significant prognostic biomarker for poor outcomes in a broad population of trauma patients. The pooled data demonstrated a consistent and statistically significant association between high D-dimer concentrations and critical endpoints, including mortality and DVT. Although significant heterogeneity was observed across the included studies, the application of robust methodologies, including random-effects modeling and subgroup analyses, strengthened the validity of these findings.

Numerous studies indicated that the risk of DVT in trauma patients ranged from 2.5% to 18.91% [22, 35-38]. Consistent with this, Zang et al., [17] reported significantly higher D-dimer levels in trauma patients with DVT than both healthy individuals and non-DVT patients, suggesting its utility as a predictive biomarker. Elevated D-dimer levels were also linked to worse prognoses, with studies correlating higher levels with increased risks of both short-term [39, 40] and long-term [12] mortality. Indeed, trauma patients with elevated D-dimer levels have significantly higher odds of mortality [41]. While the present meta-analysis confirmed these associations, revealing significant differences between case and control groups, considerable heterogeneity was observed across the studies. To investigate this, we performed subgroup analyses based on trauma type (TBI, multiple trauma, and lower extremity injuries). This approach substantially reduced heterogeneity, even eliminating it in one subgroup, indicating that the type and mechanism of trauma may contribute to heterogeneity may contribute to heterogeneity.

In the subgroup analysis of five studies on TBI patients, a statistically significant difference in mean D-dimer levels was found between case and control groups. This finding is supported by previous research linking elevated D-dimer levels to a higher risk of PHI [16, 42, 43], poor functional outcomes, and increased in-hospital, 28-day, 30-day, and 90-day mortality [39, 44, 45]. A multicenter TBI database analysis further confirmed that admission D-dimer levels correlate with poor GOS scores and six-month mortality [12]. Moreover, Chen et al., in a study with long-term follow-up (mean 2.8 years, maximum 6.9 years), demonstrated a relationship between high D-dimer levels and long-term mortality in TBI patients, reinforcing D-dimer’s role as a poor prognostic indicator. A dose-response relationship was observed, wherein higher D-dimer levels corresponded to a significantly increasing mortality risk [41]. However, study heterogeneity remains, potentially due to varying outcomes and mortality prediction durations [41]. Despite this consistent association, heterogeneity persists among studies. Potential sources include variations in the specific outcomes measured, the timing of mortality prediction, the units for reporting D-dimer, and the assay methods used (e.g., ELISA versus immunoturbidimetric assays) [41, 46]. Furthermore, confounding variables, such as age and sex, are often unaccounted for, despite older age and female sex being known to independently elevate D-dimer levels. The failure of many studies to stratify by these factors likely contributes to variability. Additional confounders include anticoagulant use (e.g., warfarin, which may lower D-dimer levels) and comorbid conditions such as malignancy, which can increase them [46, 47].

In the analysis of studies on patients with multiple traumas, a significant difference was observed in mean D-dimer scores between the case and control groups, with an SMD of 0.87 (95%CI: 0.52, 1.21). Notably, no heterogeneity was detected (I²=0%), suggesting that trauma type might be a valuable predictor of outcomes influencing the association between D-dimer and patient outcomes. This finding was supported by a 2016 multicenter retrospective study of 519 adult trauma patients, which reported higher mortality in those with elevated D-dimer levels [48]. Furthermore, elevated D-dimer levels were shown to correlate with the degree of trauma severity and tissue damage, serving as an important indicator of the ensuing inflammatory process [9].

Determining changes in D-dimer levels in patients with severe multiple traumas is crucial for predicting DIC. In one study, D-dimer levels were significantly higher in trauma patients with DIC than those without [49]. Supporting this, a 2019 prospective multicenter observational cohort study by Gall et al., which involved 940 severely injured patients (mostly with TBI), reported that D-dimer levels were seven times higher in deceased patients than in survivors [50]. This strongly indicated that the mean D-dimer level could be a powerful predictor of poor outcomes in multiple trauma patients. However, the precise mechanisms linking elevated D-dimer levels to poor outcomes, such as progressive PHI, thromboembolic complications, and multi-organ failure, remain unclear [51].

Furthermore, while the analysis of patients with lower extremity injuries [17] revealed heterogeneity, a statistically significant difference in mean D-dimer levels between case and control groups was still observed. Given its consistent predictive role for adverse outcomes in trauma patients, D-dimer levels measured at admission could serve as a valuable marker for clinical management [41]. This supports the potential for establishing specific D-dimer cut-off values to predict adverse outcomes in this patient population.

This study had several limitations. The exclusion of low-quality studies and the omission of the Google Scholar database might have contributed to heterogeneity and introduced a selection bias. Furthermore, the reliance on retrospective and observational study designs, inherent assay variability in D-dimer measurement, and potential generalizability issues were other important limitations.

The findings of this review confirmed that elevated D-dimer levels upon admission were significantly associated with adverse outcomes in trauma patients, including mortality, DIC, and DVT. Therefore, the routine evaluation of D-dimer levels upon admission could aid healthcare professionals in risk stratification and clinical decision-making. Future research is required to establish precise, validated cut-off points for D-dimer to predict specific adverse outcomes.

Declaration

Ethics approval: Not applicable.

Consent to participate: Not applicable.

Consent for publication: Not applicable.

Conflict of Interests: All authors declared that they have no conflict of interest.

Declaration of Generative AI in Scientific Writing: None.

Funding: The authors received no financial support for the research, authorship, and/or publication of this article.

Data availability: Data availability is not applicable to this article as no new data were generated or analyzed in this study.

Authors’ Contribution: ENP: Drafting of the manuscript; SZA: Drafting of the manuscript; MK: Drafting of the manuscript; FM: Acquisition of the data; MH: Study concept, designing, drafting of the manuscript, critical revision of the manuscript for important intellectual content, and supervision.

Acknowledgments: The authors appreciate the staff of the trauma center at Shahid Rajaee Hospital, Shiraz, Iran.