Our findings revealed that high D-dimer levels on admission in this study population were associated with 30-day all-cause mortality and the need for MV, and D-dimer levels above 1.66 mg/L serve as a strong independent indicator of 30-day mortality in patients with a primary admission diagnosis of AECOPD. D-dimer levels above 1.33 mg/L were also good predictors of the need for MV. This association remained after adjustment for age, sex, co-morbidities, and CRP. In addition, the values of D-dimer correlated well with the values of the CRP level.
An elevated D-dimer was associated with increased total mortality in the absence of clinically overt disease [17, 18], as well as among patients with PE and aortic dissection [19, 20]. D-dimer levels also appeared to be accurate markers of the severity of severe acute pancreatitis and increased levels of D-dimer were associated with the development of organ dysfunction [21, 22]. Moreover, elevated D-dimer is associated with higher cardiovascular mortality among patients with cardiovascular disease [23–25]. Several studies have found this relationship existed in the patients with CAP and sepsis; moreover, D-dimer levels are closely related to severity of CAP [26–31]. An admission D-dimer level less than 500 ng/mL is also associated with low risk of short-term death and major morbidity in CAP patients [32]. In addition, D-dimer levels greater than 2,000 ng/mL identify a subgroup of CAP patients who are at increased risk for in-hospital death [26]. As a result, an elevated D-dimer is an independent correlate of increased mortality across a broad variety of disease states. Many studies have demonstrated an association between elevated D-dimer and poor prognosis, but no single cutpoint has been identified which consistently optimizes the prognostic value of the biomarker. CRP levels may also reflect the severity of inflammation[33]. In accordance with these previous finding, our results showed that increased D-dimer levels were associated with adverse outcome. D-dimer level (above 1.66 mg/L) on admission was shown to be useful in predicting mortality, with a high sensitivity of 87.6% and negative predictive value of 89.1%. Moreover, if the D-dimer level was above 1.33 mg/L, a high level correlated with poor outcomes of the need for MV. Consequently, these patients may require special attention after admission. D-dimer level could be used to distinguish a subgroup of patients with increased risk for both mortality and the need for MV. The correlation between coagulation factors and adverse outcome also suggests that anticoagulant therapies may be beneficial in patients with severe AECOPD.
The mechanism underlying the elevated D-dimer levels might be complicated. Activation of the coagulation cascade is a common and early event in patients with infection [7]. Severe coagulative disorder characterized by the diffuse formation of intravascular microthrombi and activation of fibrinolysis could be the dominant cause of this phenomenon [34]. Systemic pro-inflammatory cytokine responses are mediators of atherosclerosis directly contributing to plaque rupture through local inflammation, induction of procoagulant factors, and haemodynamic changes, which predispose to ischaemia and thrombosis [35, 36]. D-dimer levels may be elevated in AECOPD patients as an inflammatory and prothrombotic host response to infection. The mechanism underlying the increases in D-dimer levels may be the formation of multiple intravascular thrombi. Furthermore, increased inflammatory response during AECOPD may increase the ongoing inflammatory process associated with atherosclerosis, resulting in escalation of atherosclerosis and coronary artery luminal narrowing possibly through processes that involve fibrinolysis releasing D-dimer into the circulation. This phenomenon thereby increases the likelihood of atheroma plaque rupture and subsequent ischemia, as well as myocardial events [37]. A persistent prothrombotic state after infection may explain the epidemiologic link between infection and higher risk of cardiovascular disease [31]. Early elevation of D-dimer, together with other inflammatory factors, may contribute to the development or to the severity of several chronic diseases, including cardiovascular disease, leading to augmented mortality [17].The prothrombotic changes associated with acute inflammation reflected by the elevated D-dimer levels suggest that undetected thromboembolism may contribute to mortality in AECOPD patients. D-dimer may represent the summation of pro-coagulant balance or genetic factors, the extent of subclinical atherosclerosis, or the presence of underlying coagulation disorders that predispose to poor prognosis. Interventions, such as aspirin and statins, with beneficial effects on resolution of the prothrombotic state and inflammation, should be investigated to improve long-term outcomes after pneumonia [38]. Our results suggest that higher D-dimer levels may provide additional prognostic information. D-dimer is a simple severity assessment tool that is slightly better than other current scoring systems in predicting adverse outcomes, such as the need for MV and therapeutic failure. Measurement of the D-dimer level may have the potential of identify a subgroup of patients with COPD who are at low risk of adverse outcomes who may be suitable for outpatient therapy. This approach may also allow the selection of patients who will need a more intensive follow-up after admission, or those patients who might benefit from an antithrombotic therapy. Further prospective multiple center studies with a large number of patients are required to establish the exact role of D-dimer as a possible novel prognostic marker in AECOPD patients and form stronger conclusions.
This study has several limitations that should be considered. First, patients with elevated D-dimer levels were not systematically assessed to detect the presence of VTE. In this retrospective study, PE or DVT was ruled out by spiral CT scanning or peripheral vein compression ultrasonography only in patients with a high pretest probability. Therefore, not all VTE episode can be excluded, because of the presence of asymptomatic thrombus at the time of D-dimer measurement. Second, the overall mortality was different from that of other reports. The D-dimer cutoff values associated with mortality in the present study may be substantially different among patients, which probably reflected the fact that in our study, patients with recent surgery, recent trauma, active cancer, or a recent history of myocardial infarction or cerebral infarction were not included. Third, in our heterogeneous population, the patients were elderly, aged 80 years and older, and co-morbidities may have affected the coagulation process. Therefore, these characteristics may limit the generalizability of our findings to this age bracket. D-dimer is also influenced by other concomitant medical conditions, such as atrial fibrillation, heart failure, peripheral artery disease and renal failure [39–42]. Moreover, in this study, we did not exclude the patients exhibiting these concomitant medical conditions, which may have reduced the specificity of D-dimer in predicting mortality. Fourth, therapeutic interventions may be highly influenced by age and empirical antibiotic therapy, which may add some bias to our main results. Finally, D-dimer possesses high heritability and this fact supports the notion that D-dimer levels reflect genetic factors [43]. This study was performed in four different hospitals and most of the patients were from North China; thus, our findings may not be completely representative of the population of other locations. Moreover, D-dimer may not have the same testing characteristics when applied to other populations, and will probably perform worse. Although caution is necessary in the interpretation of our results, we consider it improbable that these limitations have influenced our main findings.