Following calculation of QS, a cut-off of 18 points was established by receiver operating characteristic (ROC) curve analysis (Figure 1). The population was divided into two groups – group A (n=39): <18 points and group B (n=63): ≥18 points – and the clinical, laboratory, ECG, echocardiographic and MDCT angiography parameters were compared.

ROC curve for Qanadli score and troponin I for right ventricular dysfunction (right ventricular/left ventricular ratio >1).'> ROC curve for Qanadli score and troponin I for right ventricular dysfunction (right ventricular/left ventricular ratio >1).' title='ROC curve for Qanadli score and troponin I for right ventricular dysfunction (right ventricular/left ventricular ratio >1).'/>

Figure 1.

ROC curve for Qanadli score and troponin I for right ventricular dysfunction (right ventricular/left ventricular ratio >1).

(0.14MB).

MDCT images were acquired in the caudocranial direction (collimation 1.25mm) to allow accurate imaging of the main, lobar, segmental and subsegmental arteries of the lower, middle and upper lobes.

A 120ml bolus of iodinated contrast (concentration 350mg/ml) was injected into the left antecubital vein (whenever possible) at a rate of 4ml/s, using an automatic injector system.

The images were reviewed independently by two radiologists on a workstation, without knowledge of the patients’ clinical variables, diagnostic exam results or in-hospital outcome.

To define the MDCT obstruction index, the arterial tree of each lung was regarded as having 10 segmental arteries (three to the upper lobes, two to the middle lobe and to the lingula, and five to the lower lobes), each obstructed artery being scored 1 point.12 Emboli in a main or lobar artery were given a score equal to the number of non-vascularized segmental arteries. To provide additional information on residual distal perfusion, a weighting factor was assigned to each value, depending on the degree of vascular obstruction (total occlusion=2 points) (Table 1), giving a maximum score of 40 points.

Besides evaluation of clot burden by MDCT, vascular (superior VC and AV) and cardiac dimensions (RV, LV, RV/LV ratio, CS, PA, Ao, PA/Ao ratio) related to the pulmonary tree were also measured. IVS bowing and reflux of contrast into the inferior VC were assessed qualitatively to determine signs of RVD (Figures 2–5).

Figure 2.

Right ventricular/left ventricular ratio.

(0.14MB).

Figure 3.

Measurement of coronary sinus by multidetector computed tomography.

(0.15MB).

Figure 4.

Embolus in the right pulmonary artery and left lower lobe.

(0.12MB).

Figure 5.

Central pulmonary embolism.

(0.13MB).

ECG was performed in all patients as part of the initial clinical assessment on admission to the emergency department. A modified ECG score was calculated based on a previously validated system15 (Tables 2 and 3), from which certain variables were excluded, and the points attributed to each variable were adjusted. The quantification in mm of T-wave inversion in V1–V3 was not included as we considered that the presence of T-wave inversion in these leads was qualitatively assessed in the ECG score. Furthermore, a S1Q3T3 pattern was only included when all its components were present.

SPSS was used for statistical analysis of the results, which are presented as means±standard deviation and overall percentages. The variables under study were compared using the chi-square test and independent t test. ROC curves were constructed to establish the best cut-off for QS. A value of p<0.05 was considered statistically significant.

Of the 102 patients included (overall mean age 61.5 years), 36.7% made up group A and 63.3% group B (Table 4), with an overall predominance of women (63.3%), which was maintained in both groups (Figures 6 and 7). The mean clot burden score was 18.46 points.

Figure 6.

Distribution by group.

(0.06MB).

Figure 7.

Distribution by gender.

(0.07MB).

Analysis of risk factors for PE showed no differences in the percentage of patients with deep vein thrombosis. Recent surgery was more prevalent in group A, other risk factors being more common in group B, but without statistical significance.

With regard to symptoms at admission, there was a greater prevalence in group B of fatigue, chest pain and syncope/lipothymia (p=0.017), as well as higher Geneva and Wells scores and shock index (p=0.006) (Table 5).

In terms of ECG parameters, mean heart rate (HR) and percentage of right bundle branch block (RBBB), T-wave inversion V1–V3 and S1Q3T3 pattern (p=0.034) were higher in group B, reflected in higher mean modified ECG scores (Table 6).

Laboratory tests revealed that group B had higher troponin and d-dimers, with lower creatinine clearance by the MDRD formula (p=0.020) and PO2/FiO2 ratio. Echocardiography showed higher pulmonary artery systolic pressure (PSAP) in group B (Table 6).

CT angiography showed larger RV diameters and higher RV/LV ratio (p=0.002), and larger superior VC, AV and CS diameters in group B. PA diameter and PA/Ao ratio were similar.

Interventricular septal bowing and reflux of contrast into the inferior VC (p=0.001) were greater in group B, and QS >18 points was an independent predictor of RVD (RV/LV ratio >1) (OR: 10.85 [95% confidence interval 3.20-36.77]; p<0.001) (area under the curve [AUC] on ROC analysis: 0.79; p<0.001), with a sensitivity of 78.4% and specificity of 79% (Figure 1, Table 7).

In the remainder of the univariate analysis, the following were also identified as predictors of RVD: syncope/lipothymia, tachypnea, S1Q3T3 pattern, modified ECG score >5, peak troponin I >0.10ng/ml, IVS bowing, reflux on contrast into the inferior VC and CS diameter >10mm (Table 7). Comparative analysis of the ROC curves showed that QS had a greater AUC than peak troponin, whose association with RVD has been clearly established (Figure 1).

The percentage of patients receiving fibrinolytic treatment was higher in group B (p=0.049), and in-hospital mortality was similar in both groups (overall 4.9%; n=5) (Figure 8).

Figure 8.

Treatment according to Qanadli score.

(0.07MB).

Our registry showed mean obstruction of 46.15% (mean QS 18.46 points), higher than reported by Wu et al.13 and Qanadli et al.12, which can be explained by the clinical characteristics of the study population that included only intermediate or high-risk patients.

The groups were similar in terms of age, with a predominance of women in both groups. Although there were no significant differences in risk factors, there was a strong association between syncope/lipothymia and higher QS (p=0.017) related to the impact of clot burden on pulmonary pressure and RV function. Clinical scores (Geneva and Wells) were slightly higher in group B, as was the shock index (p=0.006), reflecting greater morbidity in those with QS>18 (Table 5).

A variety of ECG alterations have been associated with PE severity. The ECG score developed by Daniel et al.15 (Table 2) increases with severity of pulmonary hypertension due to PE, identifying patients with greater perfusion defects.

ECG analysis revealed higher mean HR and a slightly higher percentage of RBBB in group B. However, the most significant differences were found in percentage of T-wave inversion in V1–V3 and S1Q3T3 pattern (p=0.034) and in mean modified ECG score (p=0.009), with a higher prevalence of these dynamic changes in group B, showing a linear correlation between pulmonary clot burden and RV alterations on surface ECG (Table 6).

The results obtained in our study are evidence of the relationship between clot burden and independent laboratory predictors of PE severity, demonstrating that QS reflects the interactions between different organs. QS>18 was associated with lower creatinine clearance by the MDRD formula (p=0.02) and PO2/FiO2 ratio and higher peak troponin I (Table 6).

Various studies have claimed that the association between troponin I levels and severity of RVD in PE has a significant impact on risk and prognosis.4,18,19 ROC analysis of our results showed that QS had a greater AUC than troponin I for RVD (Figure 1), while univariate analysis revealed the highest OR for this score (Table 7), reflecting its high value for risk stratification in PE.

Echocardiographic studies showed that group B had slightly higher PASP on admission to the ICU, but without statistical significance (Table 6) (a finding also reported by Qanadli et al. and Miller et al., 12,16). In PE patients, obstruction of the pulmonary vascular tree is the main factor in increased pulmonary vascular resistance, resulting in pulmonary hypertension. However, the hemodynamic profile may depend on whether there is pre-existing pulmonary disease.12 Better correlation between mean pulmonary artery pressure and the severity of obstruction has been reported in selected patients without underlying pulmonary disease, compared with unselected patients,20 as was the case in our study. The other parameters of RVD were not included in the results due to interobserver variability and the fact that data were not available for all patients.

On CT angiography, there was a linear correlation between QS and RVD parameters (RV, superior VC, AV and CS diameters, RV/LV and PA/Ao ratios, and percentage of IVS bowing and reflux of contrast into the inferior VC); the value of clot burden as assessed by QS in indicating multiple parameters of RVD has not been described previously in the literature.

ROC curve analysis determined a cut-off of 45% pulmonary vascular obstruction for the presence of signs of RVD (with significant sensitivity and specificity), which is midway between that obtained by Qanadli et al. (40%)12 and that reported by Mastora et al. (49%).14

Univariate analysis identified the following independent factors for RVD: syncope at admission, tachypnea, S1Q3T3 pattern, ECG score >5 points, troponin I >0.10ng/ml, QS>18 points, CS diameter >10mm, IVS bowing and reflux of contrast into the inferior VC; QS showed the highest OR (Table 7).

In-hospital mortality was 4.9%, with no difference between the groups.

Although previous studies have reported a weak correlation between clot burden scores and short-term mortality in patients with PE, to our knowledge there are no studies in the literature assessing the association between clot burden at admission and long-term outcomes (12-month mortality).1

Two small retrospective studies have suggested a correlation between angiographic clot burden and short-term mortality. Wu et al.13 reported that patients with a pulmonary obstruction index of more than 60% had higher mortality, and Van der Meer et al.3 found a relationship between QS and mortality, which was not borne out by our results.