The mean age of our patients was 51.43 ± 16.88, 50 (62.5%) were females and 30(37.5%) were males .The most frequent clinical presentations for pulmonary embolism among studied patients were dyspnea (95.0%) followed by chest pain (52.5%), hemodynamic instability (40.0%), hemoptysis (31.3%), lower limb swelling and pain (15.0%), cough (12.5%), and syncope (8.8%). 50% of patients received anticoagulants, 41.2% received thrombolytics and 8.8% received catheter-directed therapy (CDT). 60 patients (75%) were survived. and the rest of them was not.
Patients were classified into two groups: NO RVD group and RVD group according to Echocardiography. Table 1 shows the association between the outcome and RVD among studied patients. There was statistically significant higher percent of non -survivors among RVD group compared to no RVD group (45.2% vs 2.6% respectively), p value < 0.001.
RVD patients had significantly lower PaO2, SPO2, and higher A-a O2 gradient compared to non-RVD patients (p value0.001).
As regard the CBC parameters, RVD group showed significantly higher mean RDW (p = 0.008) and median NLR, PLR (p = < 0.001) compared to non-RVD group. However, median MPV did not differ significantly between both groups.
In addition, RVD group had considerably higher median CRP, D dimer, and troponin levels compared to non-RVD group (p < 0.001).
The most accurate serum markers able to predict RVD were A-a O2 gradient, serum troponin, CRP, D- dimer, NLR, RDW and PLR in order as shown in table 3 and Fig. 1.
Table 4 demonstrated the predictors attributed with the occurrence of RVD by univariate logistic regression analysis. These predictors included elevated A-a O2gradient, increased RDW, elevated NLR, elevated PLR, elevated CRP, increased D-dimer, and elevated troponin levels. By multivariate logistic regression analysis, the only significant predictor was an increase in the A-a O2 gradient, with an odds ratio of 1.10 and a p-value of 0.047.
Patients were classified according to 30-day mortality into two groups: survivors and non-survivors. Table 5 shows the association between ABG, serum markers and mortality among studied patients. There was statistically significant lower mean PO2, SPO2, and higher mean A-a O2 gradient among the non-survivors. While, there was no statistically significant difference in mean RDW and mean MPV between survivors and non-survivors (16.57 ± 5.485 vs 17.59 ± 2.48, p value = 0.425), (9.01 ± 1.32 vs 9.68 ± 2.18, p value = 0.102) respectively. On the other hand, there was statistically significant higher median NLR, PLR among non-survivors compared to survivors (9.15 vs 3.00, p value < 0.001), (270.00 vs 171.50, p value = 0.001) respectively. In addition, there was statistically significant higher median CRP, D dimer, troponin among non-survivors, p value < 0.001.
Table 6 and Fig. 2 show the diagnostic ability of ABG and serum markers in prediction of 30-day mortality, the most serum markers able to predict mortality were, D- dimer, A-a gradient, serum troponin, NLR, CRP and PLR in order.
The significant predictors associated with 30 day-mortality by univariate logistic regression analysis were increase A-a O2 gradient, increase NLR, increase CRP, increase D-dimer and increase troponin level as displayed by table (7), but none of them were significant by applying multivariate regression.