Recent studies have highlighted differences in how older patients respond to high-risk pulmonary embolism (PE) and treatment. However, guidelines for PE risk stratification and treatment are not based on age, and data are lacking for older patients. We characterized the impact of age on clinical features, risk stratification, treatment, and outcomes in a sample of patients with PE in the emergency department. We performed an observational cohort study of 547 consecutive patients with PE in the emergency department from 2005 to 2011 in an urban tertiary hospital. We used bivariate proportions and multivariable logistic regression to compare clinical presentation, risk category, treatment, and outcomes in patients ≥65 years with those <65 years. The mean age was 58 ± 17 years, 276 (50%) were women, and 210 (38%) were ≥65 years. PE was more severe in patients ≥65 years (massive 14% vs 5%, submassive 48% vs 25%, and low risk 38% vs 70%, p <0.0001), with submassive PE being the most common presentation in patients ≥65 years. However, subanalysis removing natriuretic peptides from the definition of submassive PE negated this finding. Treatment with parenteral anticoagulation (88% vs 90%, p = 0.32), thrombolytic therapy (5% vs 4%, p = 0.87), and inferior vena cava filter (4% vs 4%, p = 0.73) were similar among age groups. Patients ≥65 years had higher 30-day mortality (11% vs 3%, p <0.001). In conclusion, patients ≥65 years present with more severe PE and have higher mortality, although treatment patterns were similar to younger patients. Age-specific guideline definitions of submassive PE may better identify high-risk patients.
Highlights
- •
Elderly patients present with more severe pulmonary embolism (PE).
- •
Elderly patients with PE have higher 30-day mortality than younger patients.
- •
Treatment patterns do not differ among age groups.
- •
Natriuretic peptides may not be accurate markers of PE severity in the elderly population.
- •
The guideline definition of submassive PE may need to be modified for the elderly population.
Pulmonary embolism (PE) is a leading cause of morbidity and death. Older patients have a higher incidence of PE and higher PE-related morbidity and mortality than younger patients. Guidelines recommend risk-stratifying patients with PE into low risk and submassive and massive categories based on biomarker, radiographic, echocardiographic, and electrocardiographic parameters and suggest basing triage and treatment decisions on these categories. Recent studies have identified older patients as being at higher risk from the treatment of PE, but few previous studies have focused on PE in patients ≥65 years, so guideline definitions of submassive and massive PEs do not differ for older patients. However, biomarkers and other cardiac parameters are known to change with age, and these changes may decrease the accuracy of risk stratification in patients ≥65 years. Recent studies validating age-specific biomarker interpretation support this. We, therefore, studied the impact of age on PE risk stratification, clinical features, treatment, and outcomes in patients with PE in the emergency department (ED).
Methods
We analyzed data from a single-center, observational study performed at a tertiary referral hospital with an annual ED volume of 95,000 patient visits. For the purposes of this analysis, patients from 2 cohorts were grouped. The first was a prospective cohort of 298 enrolled consecutive patients diagnosed with acute PE in the ED from October 2008 to December 2011, and the second was a retrospective cohort of 249 consecutive patients diagnosed with acute PE in the ED from May 2005 to April 2008. The same data form and definitions were used to collect data from both cohorts. Further details regarding the design and methodology of both studies have been previously published.
Adult patients (age >17 years) were eligible for this study. PE was defined as (a) a filling defect in a pulmonary artery consistent with thromboembolism on computed tomographic pulmonary angiography, (b) a high probability ventilation perfusion (V/Q) lung scan, and/or (c) a lower extremity venous duplex ultrasound or computed tomographic venogram demonstrating deep venous thrombosis in association with PE symptoms, when the treating clinician confirmed that the test was performed to evaluate suspected PE, not isolated deep vein thrombosis.
Patients were eligible for enrollment if they were diagnosed with PE within 24 hours after registering in the ED. Patients transferred from outside EDs or clinics with a diagnosis of PE were eligible for enrollment until 24 hours after their outside hospital radiographic procedure. This study complies with the ethical rules for human experimentation that are stated in the Declaration of Helsinki, including approval of an institutional review board and informed consent obtained from all enrolled patients.
We collected demographics, co-morbid illnesses, ED vital signs, laboratory results, radiological findings, ED treatment, disposition, and selected outcome measures including admission to intensive care unit, “major bleeding,” and 5-day and 30-day mortality for each enrolled patient. Biomarker (troponin-T, N-terminal prohormone of brain natriuretic peptide [NT-proBNP], d -dimer) and imaging studies were performed at the discretion of the treating physician. Only biomarkers performed within 24 hours of diagnosis were included. The results of echocardiograms performed within 3 days of diagnosis were included. We also collected blood samples from enrolled patients within 24 hours of PE diagnosis and used these to perform biomarker analysis when results were not available clinically. Using this combination of tests, >90% of subjects had a complete panel of biomarkers available for analysis. We also collected data necessary to define the Pulmonary Embolism Severity Index, which is a validated clinical score designed to identify patients at risk for 30-day all-cause mortality after acute PE. We divided our population using an age cutoff of ≥65 years, as used by the World Health Organization.
“Massive PE,” “submassive PE,” and “low-risk PE” were defined according to the guidelines of the American Heart Association. Massive PE was defined as PE with systolic arterial hypotension (systolic blood pressure <90 mm Hg) and was based on lowest recorded ED blood pressure. Submassive PE was defined as PE with normal blood pressure (systolic blood pressure ≥90 mm Hg) plus evidence of right ventricular (RV) dysfunction or “myocardial necrosis.” “RV dysfunction” was defined based on electrocardiogram (RsR in V1; right bundle branch block or ST-segment elevation in leads V1, V2, and V3; ST-segment depression in leads V1, V2, and V3; or T-wave inversion in leads V1, V2, and V3), echocardiogram (RV dilatation, hypokinesis, or bowing of the intraventricular septum), or NT-proBNP >500 pg/ml. Myocardial necrosis was defined as a troponin T >0.01 ng/ml. Low-risk PE was defined as PE with normal blood pressure and no evidence of RV dysfunction. To facilitate analysis, we also combined massive and submassive PE into a single category called “severe PE.” “Recent immobilization” was defined as a hospitalization lasting ≥3 days, surgery, leg trauma (or in a cast), or other bedbound state within the last 30 days. “Recent travel” was defined as having duration of ≥6 hours and occurring within the last 30 days. Major bleeding was defined as intracranial, gastrointestinal, or retroperitoneal bleeding or any bleeding requiring a transfusion of ≥2 units of packed red blood cells.
We followed patients daily for 5 days while in the hospital and for 30 days after PE using a validated combination of telephone calls and medical record review.
In our primary analysis, we compared clinical features, PE risk category, treatment, and outcomes across 2 age groups: patients <65 years and those ≥65 years. We present continuous variables as mean ± SD and categorical variables as percentages (%). We used Student t tests to compare continuous variables and chi-square tests to compare categorical variables. We performed both multivariable logistic regression and a propensity analysis to determine the association between age ≥65 years and PE severity, controlling for a prespecified list of potential confounders (gender, coronary artery disease, congestive heart failure, renal insufficiency, lung disease, cerebrovascular disease, active malignancy, and atrial fibrillation/flutter). NT-proBNP increases with age and including NT-proBNP in our definition of submassive PE could influence our results, so we performed a sensitivity analysis omitting NT-proBNP from the definition of submassive PE. We also performed a sensitivity analysis comparing patients ≥80 years with those <80 years. A 2-sided p value of ≤0.05 was considered as statistically significant. Analyses were conducted using SAS 9.3 (The SAS Institute, Inc., Cary, North Carolina).
Results
Overall, the cohort consisted of 547 patients, 298 (54%) from the prospective cohort and 249 (46%) from the retrospective cohort. There were no significant demographic differences between the prospective and retrospective cohorts. The cohort mean age was 58 ± 17 years. Two hundred seventy-six patients (50%) were women, three hundred thirty-seven (62%) were <65 years, 210 were ≥65 years (38%), and sixty-five patients (12%) were ≥80 years. The mean age of patients in the <65-year-old group was 47 ± 12 years; those in the ≥65-year-old group was 75 ± 7 years; and those in the ≥80-year-old group was 84 ± 4 years. Patients ≥65 years were more likely to have a history of co-morbid conditions and more likely to present with hypoxemia ( Table 1 ). Presenting clinical features were otherwise similar between groups.
| Variable | Age<65 years (N=337) | Age≥65 years (N=210) | P value |
|---|---|---|---|
| Age (years) ∗ | 47±12 | 75±7 | <.0001 |
| Female | 161 (48%) | 115 (55%) | .11 |
| White | 275 (82%) | 186 (89%) | .029 |
| Coronary artery disease | 18 (5%) | 36 (17%) | <.0001 |
| Congestive heart failure | 14 (4%) | 17 (8%) | .054 |
| Left ventricular ejection fraction <35% | 6 (2%) | 3 (1%) | .76 |
| Renal insufficiency | 1 (<1%) | 8 (4%) | .002 |
| Chronic obstructive pulmonary disease | 7 (2%) | 27 (13%) | <.0001 |
| Cerebrovascular disease | 15 (4%) | 15 (7%) | .18 |
| Prior deep vein thrombosis or pulmonary embolism † | 35 (19%) | 19 (16%) | .50 |
| Recent immobilization † | 37 (20%) | 33 (28%) | .12 |
| Recent immobilization only related to travel † | 21 (12%) | 5 (4%) | .029 |
| All malignancy | 96 (28%) | 94 (45%) | .0001 |
| Active malignancy: under treatment or metastatic | 60 (18%) | 48 (23%) | .15 |
| Highest heart rate (beats per minute) ∗ | 96±20 | 97±21 | .73 |
| Lowest systolic blood pressure (mm Hg) ∗ | 113±18 | 113±26 | .95 |
| Tachypnea (respiratory rate ≥20 breaths per minute) | 264 (78%) | 176 (84%) | .081 |
| Hypoxia (oxygen saturation <90%) | 18 (5%) | 24 (11%) | .009 |
| Atrial fibrillation/flutter on electrocardiogram | 7 (2%) | 19 (9%) | .0002 |
Patients ≥65 years had a higher prevalence of most laboratory and imaging measures of severe PE compared with patients <65 years ( Table 2 ) and accordingly were more likely to be categorized as having higher PE severity on presentation (massive 14% vs 5%, submassive 48% vs 25%, and low risk 38% vs 70%, p <0.0001; Figure 1 ). Patients ≥65 years were most likely to present with submassive PE, whereas patients <65 years were most likely to present with low-risk PE. Among all patients, 73 had submassive PE based on NT-proBNP alone (13% of enrolled and 40% of submassive PE). In patients ≥65 years, 52 had submassive PE based on NT-proBNP alone (25% of enrolled and 52% of submassive PE), whereas 21 patients <65 years had submassive PE based on NT-proBNP alone (6% of enrolled and 25% of submassive PE). In 140 patients with elevated NT-proBNP, 20 (14%) had a history of congestive heart failure and 6 (4%) had a previously documented left ventricular ejection fraction <35%. The mean Pulmonary Embolism Severity Index score, which includes age, was also higher in patients ≥65 years (119 vs 77, p ≤0.0001).
| Variable | Age<65 years (n=337) | Age≥65 years (n=210) | P value |
|---|---|---|---|
| Measures of Pulmonary Embolism Severity: | |||
| Pulmonary embolism severity index ∗ | 77±33 | 119±34 | <.0001 |
| Troponin T ≥0.1 ng/ml | 12 (4%) | 18 (9%) | .012 |
| NT-proBNP >500 pg/ml | 45 (13%) | 95 (45%) | <.0001 |
| Right ventricular strain on electrocardiogram | 38 (11%) | 30 (14%) | .30 |
| Right ventricular strain on trans-thoracic echocardiogram | 30 (9%) | 31 (15%) | .034 |
| Central pulmonary embolism on computed tomography pulmonary angiography | 155 (46%) | 101 (48%) | .63 |
| Multiple pulmonary emboli on computed tomography pulmonary angiography | 246 (73%) | 153 (73%) | .97 |
| Pulmonary Embolism Severity Category: | <.0001 | ||
| Low-risk pulmonary embolism | 237 (70%) | 79 (38%) | |
| Submassive pulmonary embolism | 84 (25%) | 101 (48%) | |
| Massive pulmonary embolism | 16 (5%) | 30 (14%) | |
| Treatment: | |||
| Anticoagulation † | 304 (90%) | 185 (88%) | .32 |
| Thrombolysis | 15 (4%) | 10 (5%) | .87 |
| Inferior vena cava filter | 15 (4%) | 8 (4%) | .73 |
| Outcomes: | |||
| Admission to intensive care unit | 22 (7%) | 22 (10%) | .099 |
| Major bleeding | 6 (2%) | 8 (4%) | .14 |
| In-hospital mortality ‡ | 0 (<1%) | 2 (1%) | .073 |
| 30-day mortality | 9 (3%) | 23 (11%) | <.0001 |
† Anticoagulation started in the emergency department.
‡ In-hospital mortality defined as within 5 days of emergency department registration.
On multivariable analysis, age ≥65 years was an independent predictor of PE severity ( Table 3 ). Similarly, on propensity analysis (matched on coronary artery disease, congestive heart failure, renal insufficiency, lung disease, cerebrovascular disease, active malignancy, atrial fibrillation/flutter), patients ≥65 years were more likely to have massive/submassive PE (p <0.001). In our subanalysis of patients with severe (the combination of massive and submassive) PE, patients ≥65 years were more likely than patients <65 years to have an elevated NT-proBNP (73% vs 45%, p <0.0001) but were not more likely to have other markers of PE severity (data not shown).
