Tissue Plasminogen Activator Plasma Levels as a Potential Diagnostic Aid in Acute Pulmonary Embolism

Pulmonary embolism (PE) is a potentially fatal and frequent complication of deep venous thrombosis (DVT).1 Its diagnosis is difficult because clinical symptoms and signs and complementary routine exploration tests (hemogram, blood gases, chest radiography, and electrocardiography) are nonspecific.2–4 On the other hand, the most reliable techniques for the diagnosis of PE, such as ventilation-perfusion (V/Q) lung scan and pulmonary angiography, are not universally available in many hospitals and have other limitations. As such, the most important screening test for the diagnosis of PE is V/Q lung scan, which is only definitive (confirming or excluding PE) in 30% to 50% of cases,5,6 and pulmonary angiography, which is a high-cost, invasive technique with notable morbidity and mortality.7 

In the last few years, some plasma determinations, such as D-dimer, have been used in the diagnosis of PE.8 It has been indicated that D-dimer performed by an enzyme-linked immunosorbent assay (ELISA) technique can exclude the diagnosis of PE if the patient's value is less than 500 ng/mL.9–13 D-dimer is one of the last products derived from the thrombotic/fibrinolytic process. Its levels rise when thrombin activates fibrinogen, fibrin is formed, and factor XIIIa stabilizes it14; then, tissue plasminogen activator (tPA) activates plasminogen to plasmin and plasmin splits fibrin, one of these fibrin degradation products being D-dimer.14,15 This fibrinolytic process is regulated by plasminogen activator inhibitor type 1 (PAI-1) inhibiting tPA in the circulation, and α₂-antiplasmin inhibiting free plasmin.16 

In addition to hypercoagulability, fibrinolytic system activation is necessary for the rise of plasma D-dimer levels. Fibrinolytic system activation means that tPA must be released into the circulation. This theoretic conclusion led us to investigate the fibrinolytic system activators, inhibitors, and complexes (tPA, PAI-1, plasmin-antiplasmin complex [PAP], and D-dimer levels) in suspected PE patients.

Sixty-six (66) consecutive outpatients who presented with clinically suspected PE in the emergency room of Hospital Príncipe de Asturias (Madrid, Spain) between January 1997 and March 1998, and who underwent lung scan, were included in the study. The University Hospital Príncipe de Asturias is both a teaching and a general hospital; it has 350 beds and serves an area with 280 000 inhabitants.

The medical history of each patient and the findings on physical examination were recorded. Chest radiography and electrocardiography were performed. On the basis of these findings, we assessed the clinical probability of PE as low, intermediate, or high.17 

A V/Q lung scan with a Anger-type camera was carried out on all the patients in 6 projections (anterior, posterior, left lateral, right lateral, left posterior oblique, and right posterior oblique) in the first 48 hours after their admittance to the hospital. Perfusion scan images were obtained after intravenous injection of 111 to 185 MBq of macroaggregate albumin marked with technetium-99m (^99mTc). Ventilation scan images were obtained with inhalation of ^99mTc pyrophosphate radioaerosol. Each patient's lung scan was interpreted as having normal, low, intermediate, or high probability for PE by 2 independent observers from the Nuclear Medicine Department; also, each patient was assessed as having a low, intermediate, or high clinical probability of having PE, as described in the Prospective Investigation of Pulmonary Embolism Diagnosis (PIOPED) study.5 A normal perfusion lung scan result ruled out the diagnosis of PE. A high-probability V/Q lung scan and a high clinical probability were definitive for confirming a diagnosis of PE.

When the clinical and lung scan results were nondiagnostic (inconclusive), a color Doppler ultrasound (CDUS) (Eco Duplex Color P 600, Philips Medical Systems, Irvine, Calif) study of the lower limbs was done to diagnose acute DVT. Color Doppler ultrasound examined the superficial and deep veins of the lower limbs, longitudinally and transversally, from the calf to the inferior vena cava. The diagnosis of acute DVT was based on the finding of direct intraluminal image or absence of complete venous compressibility. If CDUS was uncertain, a contrast venography study of the lower limbs was performed. Those patients with either a positive venography or positive CDUS were considered PE positive. If venography or CDUS was inconclusive, a contrast pulmonary angiography was performed to confirm the diagnosis. The angiograms were interpreted by a specialist in vascular radiology, who considered the presence of repletion defects or sharp termination of 1 or more arteries greater than 2.5 mm in diameter as positive for PE diagnosis.

Finally, similar to earlier studies,9,10,12,18,19 patients were classified as PE positive when 1 of the following findings was present: (1) positive pulmonary angiography, (2) high-probability V/Q lung scan and high clinical probability, or (3) inconclusive V/Q lung scan and positive lower-limb examination for DVT (CDUS or venography). Patients were classified as PE negative when 1 of the following findings were present: (1) normal perfusion lung scan, (2) very low or low-probability V/Q lung scan and low clinical probability in the absence of DVT in lower limbs (CDUS or venography), or (3) normal pulmonary angiography. A flowchart of the diagnostic process is shown in Figure 1. In all suspected PE patients, intravenous heparin therapy was started at the time of presentation and continued until objective testing could be performed. However, once the test was performed and patients were classified as PE positive or PE negative, appropriate treatment was instituted. The PE-negative patients were not treated with anticoagulant and were monitored for 3 months to determine whether they had any symptoms suggestive of DVT or PE.

Within 24 hours of presentation, blood was obtained from all patients by venipuncture in plastic tubes containing 3.8% trisodium citrate (9:1, vol:vol). Specimens were centrifuged at 3000g for 15 minutes to obtain platelet-poor plasma, which was aliquoted and stored at −70°C. The technician performing the analysis was unaware of the final diagnosis for each patient.

Tissue plasminogen activator, PAI-1, PAP, and D-dimer levels were assessed using an ELISA method with the following kits: tPA (TintElize tPA, Biopool, Umea, Sweden),20 PAI-1 (TintElize PAI, Biopool),21 PAP (Enzygnost-PAP, Bhering Institute, Marburg, Germany),22 and D-dimer (VIDAS, bioMérieux, Marcy l'Etoile, France).23 The results were compared with values from a control group of 32 healthy subjects matched for age and sex. Data are expressed as mean ± standard deviation.

The comparison of quantitative parameters was done with the Student t test for variables with a parametric distribution and the Mann-Whitney test for variables with a nonparametric distribution.

A graphic distribution of values for each parameter for PE-negative and PE-positive patients was completed and, after viewing the distribution, the optimal points were chosen for calculating sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) for each test, which were calculated by standard methods.24 

This series included 66 cases with clinical suspicion of PE. The mean age of patients was 53 years (range, 21–82 years); 36 patients were women and 30 were men. Twenty-seven patients were classified as patients with PE (PE positive; 40.9%) and 39 as patients without PE (PE negative; 59.1%). The PE-positive group included 12 men and 15 women, while the PE-negative group included 17 men and 22 women (P = .926). The age of the PE-positive patients was 56 ± 16 years (mean ± SD) and the age of PE-negative patients was 52 ± 15 years (P = .301). Table 1 summarizes the results of the diagnostic strategy used in the study. One patient was classified as PE positive on the basis of a high clinical probability, intermediate-probability lung scan, and negative CDUS. This patient was judged too ill to be subjected to pulmonary angiography. None of the 39 PE-negative patients had thromboembolic events during the 3-month follow-up.

The mean plasma levels of the fibrinolytic parameters analyzed (ie, tPA, PAI-1, PAP, and D-dimer) in the PE-positive and PE-negative groups and healthy control group are shown in Table 2. The PE-positive patients showed significantly higher levels of tPA, PAI-1, and D-dimer than the PE-negative patients. The mean plasma level of PAP was similar in PE-positive and PE-negative patients. The levels of tPA, PAI-1, D-dimer, and PAP were significantly higher in PE-positive and PE-negative patients than in the healthy control subjects.

The individual distribution of tPA values in PE-positive and PE-negative groups is shown in Figure 2. The dashed line indicates the most useful cutoff value for this parameter (8.5 ng/mL). The sensitivity was 100%; specificity, 33.3%; NPV, 100%; and PPV, 51% (Table 3). Thirteen (19.7%) of the 66 patients with clinical suspicion of PE had a value lower than 8.5 ng/mL; all of these patients belonged to the PE-negative group. All the patients in the PE-positive group had values higher than 8.5 ng/mL.

The individual distribution of PAI-1 values of both the PE-positive and PE-negative groups is shown in Figure 3. The dashed line indicates the most useful cutoff value for this parameter (15 ng/mL). Using a cutoff value of 15 ng/mL, the sensitivity was 100%; the specificity, 23%; NPV, 100%; and PPV, 47.3% (Table 3). Nine (13.6%) of the 66 patients had values lower than 15 ng/mL, all of these corresponding to the PE-negative group. The distribution of individual values of both groups for D-dimer is shown in Figure 4. Using a cutoff value of 500 ng/mL, the sensitivity was 92.6%; specificity, 35.8%; NPV, 87.5%; and PPV, 50% (Table 3). The distribution of PAP values showed an excessive overlap between the PE-positive and PE-negative groups, and it was impossible to choose a cutoff point with diagnostic utility (data not shown).

In this study, a cutoff value of 8.5 ng/mL in the plasma level of tPA was the most efficacious test for excluding PE, its sensitivity and NPV being 100%. In 13 (19.7%) of the 66 patients, an exclusion diagnosis could be made without any other diagnostic test. Plasminogen activator inhibitor type 1 levels also showed 100% sensitivity and an NPV of 100%, but this test was efficacious in only 9 (13.6%) of the 66 patients.

Some previous studies have evaluated the efficacy of D-dimer determination in PE exclusion diagnosis, but we know of no study that has analyzed tPA, PAI-1, and D-dimer levels together. Speiser et al25 analyzed tPA and PAI-1 levels in patients with suspected DVT and showed lower mean levels than determined in our cases, without any difference between positive and negative DVT patients. In that study, however, only DVT patients were included, some of them with up to 14 days of evolution, and no PE patients were included. We know that venous systemic endothelium releases some tPA after vascular occlusion by the thrombotic process, thus activating the fibrinolytic system.15 However, we believe that a higher elevation of tPA level in PE patients is due to the release of tPA by the pulmonary arterial endothelium to cope with the thrombus, which could explain the elevated plasma level of tPA in this study.

The presence of false-positive cases using a tPA cutoff level of 8.5 ng/mL can be easily justified due to the large number of clinical situations and stimuli that can release tPA from the endothelium.15 Among these are nonthrombotic arterial occlusion, catecholamines, bradykinins, histamine, and others; many of these agents are released in the inflammatory reaction caused by diseases simulating PE, such as ischemic cardiopathy, pneumonia, pericarditis, or bronchial asthma. We observed similar pathologic conditions in the 39 PE-negative patients, for whom the most frequent diagnoses were pneumonia (20%), acute bronchitis with bronchial hyperreactivity (20%), myocardial ischemia and pericarditis (13%), congestive heart failure (5%), bronchial asthma (5%), and exacerbation of chronic obstructive pulmonary disease (5%) (data not shown). In the remaining 23% of PE-negative patients, an alternative diagnosis could not be made. The presence of false-negative cases seems unlikely, because tPA deficit is an exceptional anomaly in the fibrinolytic system.15,26 

Although a PAI-1 cutoff value of 15 ng/mL showed a sensitivity and NPV of 100% in the diagnosis of PE, a low (13.6%) number of true-negative patients was found. These results show that the clinical utility of PAI-1 was lower than that of tPA. The high number of false-positive cases is not surprising, due to the fact that PAI-1 can be considered an acute-reactant-phase parameter, its level being elevated in ischemic cardiopathy and sepsis.15,16 It has been demonstrated that some hormones, growth factors, and cytokines implicated in inflammatory and metabolic reactions can stimulate the synthesis and release of PAI-1 from the endothelial cells.15 

Many studies seem to have demonstrated the efficacy of D-dimer level determination by an ELISA method to rule out PE,9–13 but its practical diagnostic utility has been limited by a lack of standardization, high cost, and technical difficulties in individual cases.27,28 

Results of a new rapid, individual, automated quantitative D-dimer ELISA have recently been published (VIDAS D-dimer).23 This test seems to have a sensitivity and NPV of 100% when a cutoff value of 500 ng/mL is chosen.23 In the present study, the sensitivity and NPV values obtained with rapid VIDAS D-dimer were 92.6% and 87.5%, respectively, which are slightly low compared with values reported in previous studies, but are similar to sensitivity values and NPVs obtained in recent studies using the same test.29 This discrepancy could in part be due to a possible confounding effect of heparin treatment during the diagnostic workup. Minnema et al30 indicated that heparin therapy lasting more than 24 hours increases plasma levels of tPA and decreases plasma levels of D-dimer in patients with or without PE. In our study, the possible confounding effects of heparin therapy is very unlikely, because the plasma samples were obtained less than 24 hours after the initiation of heparin therapy, and heparin therapy was started and maintained in all suspected PE patients (both PE positive and PE negative) until objective diagnostic tests were completed. However, other investigators have not observed clear effects of heparinization on plasma levels of fibrinolytic biomarkers such as D-dimer.31,32 Recently, Heit et al33 observed that plasma D-dimer sensitivity is unaffected by intensity or duration of heparin treatment in patients with suspected PE. Therefore, new studies are needed to clarify this question. Using the VIDAS D-dimer we observed 2 false-negative cases, corresponding to cases with intermediate-probability lung scan and intermediate clinical probability for PE and DVT in the lower limbs demonstrated by contrast venography. Because tPA release is necessary for a rise in D-dimer, it has been said that a tPA release deficit could explain some false-negative cases.26 Our 2 patients, however, showed high levels of tPA (31.9 and 21.4 ng/mL). Another explanation could be that excessively elevated plasma levels of PAI-1 could inhibit tPA level in the circulation, thus inhibiting the production of D-dimer34; this explanation would be the most likely, because the 2 false-negative cases had high plasma levels of PAI-1 (36.3 and 35 ng/mL).

In conclusion, even though the number of patients in our study was not very high, our data indicate that plasma tPA and PAI levels are potentially useful for the exclusion of PE in outpatients; tPA seems to be the better parameter of the two. The rapid D-dimer ELISA showed lower NPV and sensitivity values than reported in previous studies using the same test.

We acknowledge Mariano Pérez Pascual for his technical assistance.