This retrospective cohort study in a tertiary care hospital included consecutive patients admitted with HR-PE over a 10-year period, using International Classification of Diseases (ICD)-9 codes for the period between January 1, 2008 and December 31, 2015 and ICD-10 for patients admitted between January 1, 2016 and December 31, 2018.

Patients who fulfilled HR-PE criteria according to the most recent European Society of Cardiology/European Respiratory Society (ESC/ERS) guidelines were included.1

Hemodynamically stable patients (low- or intermediate-risk PE) were excluded from the study, including those who developed hemodynamic instability after the initial PE diagnosis. Patients with a presumptive diagnosis (diagnosis not confirmed after imaging techniques or autopsy) were also excluded.

The incidence of HR-PE was calculated using the total population of 350000 that is estimated to be served by our hospital.

Data on demographic characteristics, comorbidities, clinical presentation, treatment, and clinical course were collected from clinical records. Based on these data, we calculated prognostic stratification scores that have been validated for less severe forms of PE (PESI [pulmonary embolism severity index] and sPESI [simplified PESI]).1 Data on events and mortality during follow-up were collected from hospital clinical records (emergency room or hospital admissions), as well as other hospital or primary care appointments available from a national electronic registry (Registo de Saúde Electrónico).

All patients were followed until the last available record or date of death.

The study protocol was approved by the ethics committee and was in accordance with the 1975 World Medical Association's Declaration of Helsinki. The corresponding authors assume full responsibility for the integrity of data and data analysis.

The primary objective of this study was to estimate the reperfusion rate in the HR-PE population. Reperfusion was defined as systemic or catheter-directed fibrinolysis, surgical embolectomy, or percutaneous catheter-directed embolectomy.

Secondary objectives were to identify predictors of non-reperfusion and to determine the association of non-reperfusion with adverse events: in-hospital mortality, early mortality (at 30 days) and long-term mortality and MACE (at one year and during follow-up).

Thus, all-cause mortality was established as the primary endpoint. The secondary endpoint was defined as MACE – a composite of cardiovascular mortality, embolic recurrence and occurrence of chronic thromboembolic pulmonary hypertension (CTEPH) – during follow-up.

A safety endpoint was also defined as the occurrence of bleeding events, stratified by their severity according to the Bleeding Academic Research Consortium (BARC) classification.11

Respiratory exhaustion was defined as bradypnea (respiratory frequency <12 cycles/min [cpm]), marked polypnea (respiratory frequency >30 cpm), or need for endotracheal intubation.

Recurrent PE was defined as hospital admission for a new PE event or evidence of new symptomatic PE documented on computed tomography or perfusion defects on pulmonary scintigraphy, during a follow-up appointment.

CTEPH was defined according to the 2015 European guidelines on pulmonary hypertension (with a definite diagnosis of pulmonary hypertension based on measurements obtained during right heart catheterization).12

Continuous variables were tested for normality and equality of variances using the Shapiro–Wilk test and the Levene test, respectively. Those that followed a normal distribution were presented as mean±standard deviation; those that followed a non-normal distribution were displayed as median±interquartile range. According to their distribution, groups were compared with the independent Student's t test or the Mann–Whitney test, respectively.

Categorical variables were presented as numbers and percentages and groups were compared using the chi-square test or Fisher's exact test, as appropriate.

Independent predictors of non-reperfusion were determined by multivariate logistic regression analysis. Variables inserted into the multivariate model were selected according to their significance in univariate testing (those with a p-value <0.1 and narrow 95% confidence intervals were included). The final model was built by selecting forward stepwise variables with entry criterion p=0.05 and exit criterion p=0.1. The model's goodness of fit was determined by calculating the Hosmer–Lemeshow statistic.

The cumulative incidence of events was estimated by the Kaplan–Meier method and their difference determined by the log-rank test. The benefit of reperfusion for long-term events was adjusted to the population characteristics using stepwise Cox regression analysis.

All reported p-values were two-sided and a p-value <0.05 was considered statistically significant.

The statistical analysis was performed with IBM SPSS Statistics 21.0 (IBM Corp, Armonk, NY, USA).

Between January 1, 2008 and December 31, 2018, 1955 patients were admitted to our hospital with a diagnosis of PE, resulting in an annual incidence of 58/100000 population. Of these, 74 patients (3.8%) were hemodynamically unstable at the time of the initial diagnosis, thus being considered as HR-PE and fulfilling the inclusion criteria for this study. Of the included patients, 73% (n=54) came from the emergency department (ED) and the other 27% (n=20) were already hospitalized for other reasons (Figure 1).

Figure 1.

Study flow chart. PE: pulmonary embolism.

(0,2MB).

Patients’ demographic, clinical and laboratory characteristics are shown in Table 1. Mean age was 68±15 years and most patients were female (n=49, 66.2%). The prevalence of cardiovascular risk factors such as hypertension, dyslipidemia, diabetes and overweight/obesity was very high. This population showed a higher incidence of strongly predisposing PE risk factors as defined by the ESC/ERS guidelines1 (no factor 12.2% vs. weak factor 21.6% vs. moderate factor 28.4% vs. strong factor 37.8%). The mean PESI score was 201.3±38.4. Clinical presentation was characterized by severe hypotension and the presence of acidemia (mean pH 7.3±0.2), hyperlactacidemia (mean blood lactates 6.6±6.1 mmol/l) and kidney failure (mean blood creatinine 4.8±15.5 mg/ml). A significant proportion of the patients presented respiratory failure with hypoxemia (mean partial pressure of arterial oxygen/fraction of inspired oxygen [PaO2/FiO2] 200.4±163.3) and respiratory exhaustion was present in 25.7%.

Comparison of clinical severity between ED admission and hospitalized patients showed no statistically significant differences (Table 2).

The total reperfusion rate was 50% (n=37): 35 patients underwent systemic fibrinolysis and two received CDT (one as a primary strategy and one as bailout after fibrinolysis failure). The reperfusion rate over the years and its impact on hospital mortality are presented in Figure 2. Of the 37 patients who did not receive reperfusion, 33 (89.2%) received full-dose parenteral anticoagulation: 32.4% (n=12) with unfractionated heparin and 56.8% (n=21) with low molecular weight heparin. Only four patients did not receive anticoagulation.

Figure 2.

Analysis of reperfusion rate over the years and its impact on in-hospital mortality in high-risk pulmonary embolism (PE) patients.

(0,24MB).

Analysis of variables associated with non-reperfusion showed that those that were statistically significant were older age (78±10 years vs. 61±19 years, p=0.02), previous history of coronary artery disease (87.5% vs. 45.5%, p=0.025), recent major surgery (86.7% vs. 40.7%, p=0.001) and the presence of absolute contraindications for fibrinolysis (87% vs. 45.9%, p<0.001). Some variables tended to be associated with non-reperfusion: previous cerebrovascular disease (75% vs. 45.2%, p=0.058), previous cancer (65.2% vs. 43.1%, p=0.079) and a higher mean PESI score (220.5±27.0 vs. 189.1±37.0 points, p=0.071). Moreover, although not reaching statistical significance, a higher rate of non-reperfusion was seen in patients with stronger predisposing factors (no factor 22.2% vs. weak factor 43.8% vs. moderate factor 52.4% vs. strong factor 60.7%, p=0.227).

Associated with a higher reperfusion rate were ED admission (80% vs. 38.9%, p=0.002), acidemia (pH 7.32±0.2 vs. 7.39±0.2, p=0.031) and hyperlactacidemia (6.85±1.1 vs. 3.59±2.1 mmol/l, p=0.022). Other variables including systolic blood pressure (SBP), shock index (ratio between heart rate and SBP) and PaO2/FiO2 were not associated with the type of therapy delivered.

After univariate analysis, variables that showed a statistically significant association with non-reperfusion were included in multivariate analysis (Table 3). Of the independent predictors for non-reperfusion, age over 75 years presented 12 times the risk of not receiving reperfusion therapy (reperfusion rate of 39.5% vs. 68.8%, p=0.004). Absolute contraindications for fibrinolysis were present in 31.1% of patients (Table 4) and were also independent predictors of non-reperfusion (reperfusion rate 18.5% vs. 72.0%, p<0.001): major surgery (13 patients) and central nervous system disease (eight patients) were the most frequent situations. This clinical scenario was more prevalent in previously hospitalized patients (60.9% vs. 11.8%, p<0.001), which in turn was also an independent predictor of non-reperfusion (39.3% vs. 64.9%, p=0.020).

In-hospital mortality was 36.5%. Although not reaching statistical significance, the reperfusion group presented 18.1% lower in-hospital mortality (27.8% vs. 45.9%, p=0.108), while the non-reperfusion group presented an increased relative risk for in-hospital mortality of 1.65 (95% confidence interval [CI] 0.88–3.11). The impact of reperfusion on hospital mortality is presented in Figure 2.

PE-related mortality at 30 days was 28.4% and reperfusion therapy presented a clinical benefit (17.1% vs. 40.5%, hazard ratio [HR] 0.39, 95% CI 0.2–1.0, p=0.052) (Figure 3).

Figure 3.

Impact of reperfusion on pulmonary embolism (PE)-related mortality at 30 days.

(0,09MB).

Follow-up at one year was available for 98.6% of the patients and one-year survival was 48.6%. Mean follow-up was 2.5±3.3 years, with overall survival of 33.8% (94.1% of patients had complete follow-up at 2.5 years). Of the 46 patients who survived to hospital discharge, there were seven cases of recurrent PE (15.2%) and three patients presented CTEPH (6.5%). The reperfusion benefit was more pronounced for long-term mortality, at 12 months (33.3% vs. 67.6%, HR 0.42, 95% CI 0.2–0.8, p=0.013) and during follow-up (51.4% vs. 81.1%, HR 0.42, 95% CI 0.2–0.8, p=0.005). Reperfusion also had a positive impact on one-year MACE (42.9% vs. 67.6%, HR 0.51, 95% CI 0.3–1.0, p=0.034) and MACE during follow-up (59.5% vs. 86.5%, HR 0.44, 95% CI 0.2–0.8, p=0.005). Long-term survival curves are displayed in Figures 4 and 5.

Figure 4.

Impact of reperfusion on mortality and major adverse cardiovascular events (MACE) at one year.

(0,16MB).

Figure 5.

Impact of reperfusion on mortality and major adverse cardiovascular events (MACE) during follow-up.

(0,17MB).

Most patients did not develop bleeding complications during hospital stay. There were 19 bleeding events (25.6%) (no information is available regarding 12 patients, either because they suffered very early death or because they were transferred to a different hospital). The 19 events were stratified in severity using the BARC classification: four were type 1 (no need for medical intervention), eight were type 2 (requiring medical evaluation/intervention), one was type 3a (decrease in hemoglobin of 3–5 g/dl), three were type 3b (need for surgical control of bleeding or vasoactive drugs), and three were type 5 (fatal bleeding): one central nervous system bleed and one hemoperitoneum in the non-reperfusion group and one central nervous system bleed in the reperfusion group. The reperfusion group showed a statistical tendency for higher-risk bleeding events compared to the non-reperfusion group: type 1 (6.1% vs. 6.9%); type 2 (24.2% vs. 0.0%); type 3a (3.0% vs. 0.0%); type 3b (6.1% vs. 3.4%); and type 5 (3.0% vs. 6.9%) (p=0.074). Overall, types 3c and 5 resulted in in-hospital mortality of 20% and 100%, respectively.

There are several limitations to the present study that need to be mentioned.

Firstly, this is a single-center study, so its conclusions cannot be extrapolated to the national level.

Secondly, it is a retrospective observational cohort study, which leads to several limitations including the lack of treatment randomization, and the positive impact of reperfusion may also have been influenced by differences in baseline characteristics of patients in the two groups; the small sample size may also have limited the accuracy of the results. Furthermore, although the variables used to estimate disease severity (such as hemodynamic parameters and the PESI score) and predictors of reperfusion had high reporting rates, there are some missing variables in this study, notably data from laboratory tests that are not performed in the emergency department (such as N-terminal pro-brain natriuretic peptide). Reporting bias may have led to lower estimates of event rates (recurrent asymptomatic thromboembolic events or bleeding complications). Two foreign patients were lost to follow-up at 12 months and nearly 6% of patients had no clinical records after 2.5 years.

Finally, patients with sudden death due to PE before a definitive diagnosis could not be included in this registry, which may have led to overestimation of the reperfusion rate and underestimation of mortality.

Despite these limitations, this is the first study to report the reperfusion rate and its prognostic impact in HR-PE patients in a Portuguese tertiary center. Data were collected over eleven recent years and almost 100% of 12-month follow-up and 94% at the mean follow-up at 2.5 years were achieved, which are also considered strengths of the study.