ORIGINAL ARTICLE
Nonsevere Acute Pulmonary Embolism: Prognostic CT Pulmonary Angiography Findings Mehmet Mahir Atasoy, MD,* Nesrin Sariman, MD,† Ender Levent, MD,† Rahmi Çubuk, MD,* Ömer Çelik, MD,‡ Attila Saygi, MD,† Işil Atasoy, MD,§ and Sinan Şahin, MD║ Purpose: To retrospectively evaluate the prognostic parameters of computed tomography (CT) pulmonary angiographic findings in nonsevere (hemodynamically stable) pulmonary embolism (PE) patients and to assess the predictive value of these parameters for mortality within 1 month of the initial diagnosis. Materials and Methods: Retrospectively, 67 consecutive patients (28 men, 39 women; mean age, 63.25 ± 18 years) from 2 centers with nonsevere PE diagnosed using CT and a clinical evaluation were included in the current study. Using consensus reading, 2 readers blinded to the patients' clinical outcomes quantified the right ventricle short axis to left ventricle short axis ratio in the axial plane, vascular measurements, reflux of contrast medium into the inferior vena cava and azygos vein, ventricular septal bowing, and clot load using the Qanadli scoring system. The Simplified Pulmonary Embolism Severity Index (sPESI) and pulmonary parenchymal findings were also evaluated. All CT pulmonary angiographic parameters were compared with the risk of death within 1 month using logistic regression analysis. Results: Fifty-nine patients survived (88.1%), and 8 patients (11.9%) died because of PE. The sPESI and 2 parenchymal findings (multiple wedgeshaped opacities or consolidation accompanied by a wedge-shaped opacity) were significantly related to mortality. In the univariate analysis, neither the cardiovascular CT parameters nor the clot burden was significant between the survivors and nonsurvivors (P > 0.05). Conclusions: In clinically nonsevere PE patients, the sPESI and significant parenchymal findings were the CT parameters related to 1-month mortality. Key Words: pulmonary embolism, nonsevere, hemodynamically stable, computerized tomography pulmonary angiography, prognostic findings (J Comput Assist Tomogr 2015;39: 166–170)
P
ulmonary embolism (PE) is the third most common acute cardiovascular disease.1,2 The results of large randomized trials of thrombolytics3 have indicated that clinical presentation can be considered the most powerful predictor of death caused by PE. A high mortality rate of 50% to 58% has been found in PE patients presenting with hemodynamic instability, and intensive treatment options such as fibrinolytic treatments or mechanical thrombectomy can be provided to these patients.4,5 Although nonsevere (hemodynamically stable) PE patients have a lower but still considerably high mortality rate (8%–15%), they are generally not given intensive treatments.3,4 The mortality rate for a From the *Departments of Radiology and †Respiratory Disease, Maltepe University School of Medicine; ‡Department of Cardiology, Mehmet Akif Ersoy Thoracic and Cardiovascular Surgery Training and Research Hospital; and §Departments of Cardiology and ║Radiology, Dr. Siyami Ersek Thoracic and Cardiovascular Surgery Training and Research Hospital, Istanbul, Turkey. Received for publication August 14, 2014; accepted November 19, 2014. Reprints: Mehmet Mahir Atasoy, MD, Radiology Department, Medical Faculty, Maltepe University, Feyzullah St No. 39 Postal Code 34844 Maltepe, Istanbul, Turkey (e‐mail: [email protected]). This article is an original research. It is not presented at any other congress. There is no funding received. The authors declare no conflict of interest. Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
166
www.jcat.org
subgroup of nonsevere PE patients may be reduced by providing intensive treatment. As computed tomography pulmonary angiography (CTPA) is now the first-line method for diagnosing PE,6,7 evaluating CTPA findings as possible predictors of mortality may be helpful in determining the appropriate subgroup for intensive treatment. Many studies have investigated the prognostic significance of cardiovascular CT parameters and their relationship with patient outcome, but they have not controlled for clinical presentation variables.8–14 Most previous studies have been retrospective studies using radiological data, and few studies have differentiated between severe and nonsevere PE patients and evaluated CT findings as predictors of mortality.15,16 The current study specifically analyzes nonsevere PE patients using a multidisciplinary approach. The objectives of this study were to evaluate the prognostic parameters of CTPA findings in nonsevere acute PE patients with various criteria and to assess the predictive value of these parameters for mortality within 1 month of the initial diagnosis.
MATERIALS AND METHODS Patients Retrospectively, 72 consecutive patients diagnosed as having PE between October 1, 2008, and December 31, 2012, at 2 centers were initially included in our study. Patients with hemodynamic instability were excluded. Blood pressure was taken on admission in all cases, and only hemodynamically stable patients (those with nonsevere PE) were assessed when completing the clinical probability score forms. Patients with chronic PE, determined by CT interpretation and patient history, were also excluded. Patients diagnosed (based on clinical assessment) as having pneumonia, sepsis, diffuse fibrosis, and hemoptysis (caused by a pulmonary hemorrhage) were also excluded because wedge-shaped opacities and consolidation of these diseases can cause misinterpretations of the parenchymal findings. Of the 72 nonsevere PE patients, 5 were excluded from this study (Fig. 1); the final study group included 67 hemodynamically stable PE patients. The institutional review board of our hospital approved this study, with waiver of informed consent. For each patient, the following information was collected: (a) blood pressure values at admission, (b) clinical probability assessment at admission using the Wells Scoring System, (c) patient outcome (survival or death within a 1-month period), and (d) the presence of comorbidity factors (age, presence of cancer or chronic obstructive pulmonary disease [COPD]). Patients with shock and hypotension were considered hemodynamically unstable (severe PE). “Hypotension” was defined as a systolic blood pressure less than 90 mm Hg or a pressure drop of 40 mm Hg or more for longer than 15 minutes not caused by new-onset arrhythmia, hypovolemia, or sepsis.17 Patients were accepted as nonsevere if they did not present with shock or marked hypotension. J Comput Assist Tomogr • Volume 39, Number 2, March/April 2015
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
J Comput Assist Tomogr • Volume 39, Number 2, March/April 2015
Nonsevere Acute Pulmonary Embolism
Cardiovascular CTPA Findings and Clot Burden
FIGURE 1. Flow chart of the non-severe pulmonary embolism patients.
Simplified Pulmonary Embolism Severity Index and Comorbidities Charts were reviewed for the age of the patients and the presence of either malignancy or COPD as a comorbidity. Patient outcome and mortality were retrospectively analyzed on the basis of medical charts and by using the Social Security System to establish whether patients were alive or had died during the follow-up period. The Simplified Pulmonary Embolism Severity Index (sPESI), which is a validated index for estimating 1-month mortality in patients with acute PE, was also calculated for all patients. The test is composed of the following 6 variables, which were weighted equally (1 point per variable): age older than 80 years, history of chronic cardiopulmonary disease, history of cancer, heart rate of 110 or more beats per minute, systolic blood pressure of less than 100 mm Hg, and arterial oxyhemoglobin saturation level of less than 90%. Patients with an sPESI of 1 or higher were classified as high risk and reportedly have a 1-month mortality rate of 10.9%.18
CTPA Technique and Interpretation Twenty-seven CTPA examinations were performed using a 16-detector row CT scanner (Siemens Sensation 16; Siemens AG, Medical Solutions, Erlangen, Germany), and 40 CTexaminations were performed using a 64-MDCT scanner (Aquillion 64; Toshiba Japan); images were acquired in a caudocranial direction. Patients were scanned at 120 kVp and 150 to 250 mA, according to subject body weight, with 1-mm thick sections. The images were reconstructed with a 0.8/1-mm reconstruction interval using a standard reconstruction algorithm. For intravenous access, the antecubital vein and an 18- or 20-gauge catheter were preferred. A bolus of 100 mL of the iodine contrast agent iobitridol (Xenetix 350, 350 mg/mL; Guerbet, Villepinte, France) was injected intravenously via an antecubital vein at a flow rate of 4 mL/s using a dual-head power injector, followed by a 25-mL saline chaser (5 mL/s). Bolus timing was used to optimize pulmonary artery (PA) opacification for all patients. Two radiologists with 8 and 12 years of experience in thoracic radiology interpreted the CT images. Data were transmitted to Workstation Vitrea 2 software (Vital Images Inc, Plymouth, Minn), and the images were reconstructed using postprocessing methods. Both radiologists were blinded to the clinical outcome and evaluated the CT findings by consensus reading. The CT parameters included the right ventricle (RV) short axis–to–left ventricle (LV) short axis (RV/LV) ratio in the axial plane, main PA diameter, azygos vein diameter, superior vena cava (SVC) diameter, and reflux of contrast medium into the inferior vena cava (IVC). The presence of ventricular septal bowing (VSB), pulmonary parenchymal findings, and PA clot load scores were also recorded.
The ventricular diameter measurements were obtained by measuring the short axes of the RV and LV in the axial plane at their widest points. An RV/LV axial diameter ratio cutoff of 1 was used, as recommended in the literature.19,20 We measured the transverse diameter of the main PA (through the midpoint of the pulmonary trunk) and the transverse diameter of the SVC at the level of the azygos vein. We also measured the azygos vein at its widest segment and the IVC just superior to the level of the intrahepatic vein. The IVC diameter was only measured using ultrasonography in previous studies.21,22 Because the IVC rarely appears round on CT images, we calculated the total IVC diameter as the sum of the longest diameter and the short diameter perpendicular to the longest. The PE clot load index was calculated using the method described by Qanadli et al.23 An index higher than 40% was considered to indicate a high embolic burden, which is consistent with previous literature.16 We evaluated the IVC for the presence of reflux. We accepted the VSB as positive when there was both septal flattening and septum deviation convex toward the left ventricle; VSB was considered negative in cases of septum deviation convex toward the right ventricle (normal position).
Pulmonary CT Findings The parenchymal findings were evaluated to check for wedge-shaped opacities (likely pulmonary infarction) and consolidation. A previous report suggested that these 2 parenchymal findings were significantly related to PE.24 Patients with both a wedge-shaped opacity and consolidation and patients with wedge-shaped opacities in multiple locations were considered to have significant parenchymal findings.
Statistical Analysis The NCSS (Number Cruncher Statistical System) 2007 and PASS (Power Analysis and Sample Size) 2008 Statistical Software (Kaysville, Utah) programs were used for statistical analysis. Descriptive statistics (mean, SD, median, frequency, and ratio) were calculated for the patient age, clot load scores, and all CT parameters. Student t test was used to analyze normal distributions, and the Mann-Whitney U test was used for non-normal distributions. Fisher exact test was used for the nominal variables. Pearson correlation analyses were used to determine the relationships between the parameters. The Enter method of logistic regression was used for multivariate analysis. For all tests, a value of P < 0.05 was considered statistically significant.
RESULTS Patients and Clinical Findings All 67 hemodynamically stable patients (28 men, 39 women; mean age, 63.25 ± 18 years) had blood pressures at admission higher than shock levels (range between 170/110 and 90/60). Clinical probability scores at admission were recorded according to the Wells scoring system. The Wells scores varied between 2.5 and 8, and the mean was 5.15 ± 1.59. There were no significant differences in mortality with respect to the clinical probability scores (P > 0.05) according to Fisher exact test. All patients without severe renal dysfunction were initially given low-molecular-weight heparin at weight-adjusted doses, and they continued to receive anticoagulant treatment. Eight patients were referred to the intensive care unit after admission to the hospital. Intravenous heparin was administered to 32 patients (89%). Systemic thrombolysis was performed in only 1 patient. In addition, an IVC filter was inserted in 1 patient (3%). None of the patients underwent mechanical thrombolysis.
© 2015 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
www.jcat.org
167
J Comput Assist Tomogr • Volume 39, Number 2, March/April 2015
Atasoy et al
Nine patients (13.2%) died during the 1-month follow-up period. Death was caused by malignancy in only 1 case and by PE in the other 8 patients (11.8%) according to the independent adjudication committee. The 8 patients who died as a result of PE were considered the nonsurvivor group for the statistical analysis. The average time of death was 14 days after CTPA examination (range, 2–29 days; median, 15 days). With respect to age, 18 patients were older than 70 years (26.8%), and 7 were older than 80 years (10.4%). Five patients were diagnosed as having malignancies (7.4%), and 4 patients were diagnosed as having COPD (9.0%). All patients were diagnosed as having nonsevere PE. Only 2 patients had a heart rate of 110 or more beats per minute, and 3 had a systolic blood pressure less than 100 mm Hg. Three patients had arterial oxyhemoglobin saturation levels less than 90%. There were 24 patients with sPESI of 1 or higher, with 7 belonging to the nonsurvivor group. There were no significant differences between any of the comorbidity factors and the sPESI (Fisher exact test; Table 1).
Table 1. In patients with clinically nonsevere PE, the only CT parameter that predicted mortality was the presence of significant parenchymal findings. Neither the PA CTOI nor any of the other cardiovascular parameters was correlated with the presence of significant parenchymal findings.
DISCUSSION The results of the current study demonstrate that neither cardiovascular CT parameters nor clot load scores are associated with death within 1 month in nonsevere PE patients. However, there is a strong association between the sPESI and significant parenchymal findings, which are known to be somewhat specific to PE. A number of studies evaluating CTPA have attempted to stratify patients without considering clinical presentation.8,9,13,16,25,26 Only one selected a group of patients with severe PE,15 and another evaluated patients with clinically nonsevere PE.16 In the current study, the presence of both a wedge-shaped opacity and consolidation or of multiple wedge-shaped opacities is a predictor of mortality in nonsevere PE patients. In addition, the sPESI was associated with death in nonsevere PE patients within the first month after diagnosis; this finding is consistent with previously conducted nonselective studies. The first month is a critical time for PE patients, and it is the common follow-up period. In addition, 81% of patients have been reported to show complete resolution of PE on CTPA images within 28 days of diagnosis.27 To date, many studies have reported different predictors of mortality, but the actual relationship of these characteristics with mortality is debated in the literature. Arterial blood gas analysis effectively predicts the prognosis for acute PE.28 Although parenchymal disease provides information on hemodynamic tolerance to RV dysfunction, it was not previously studied as a cause of mortality. In the current study, we did not evaluate the nonspecific findings of pulmonary parenchyma; instead, we evaluated only wedge-shaped opacities and consolidation. These 2 parameters are somewhat specific findings to PE, particularly in the absence of pneumonia, sepsis, diffuse fibrosis, and hemoptysis (caused by a pulmonary hemorrhage).23 Because we only included patients with clinically nonsevere PE in the present study, a comparison of our results with the majority of similar studies in the literature (in which the clinical status at admission is not considered) is difficult. The current study used the axial RV/LV diameter ratio, which is reportedly no less accurate than the reformatted 4-chamber RV/LV diameter ratio for predicting 1-month mortality after PE.29 We did not find a
Cardiovascular CT Measurements and Clot Burden The Qanadli scores were calculated to be between 3% and 95% with a mean of 43.99% ± 18.14%. Survivors had a mean Pulmonary Artery CT Obstruction Index (PA CTOI) of 44.41%, and nonsurvivors had a mean PA CTOI of 40.94% (P > 0.05). There was no significant difference in the mortality rate between high and low embolic burden patients (P = 0.44; Mann-Whitney U test). The RV/LV ratio in the axial plane, main PA diameter, azygos vein diameter, and IVC and SVC diameters were evaluated in all patients. There were no significant differences between the survivors and nonsurvivors with respect to these factors (MannWhitney U test). There were no significant differences in the RV/LV ratios according to the cutoff value and mortality as determined by Fisher exact test (P = 1.000; P > 0.05 for both). The cardiovascular CT parameters are listed in Table 2, along with the other quantitative measurements. Twenty-one patients (35.6%) had IVC reflux. However, this reflux was not predictive of mortality. Thirty patients (50.8%) showed VSB. We interpreted a smooth interventricular septum as VSB; the presence of VSB was not related to mortality in clinically nonsevere PE patients.
Pulmonary Parenchymal CT Findings Whereas 7 patients (7/8, 87.5%) in the nonsurvivor group had significant parenchymal findings, only 17 patients (17/59, 28.8%) in the survivor group had significant parenchymal findings (P = 0.002) (Fig. 2). The odds ratio was 17.92, as shown in
TABLE 1. Comorbidity Factors and Parenchymal Findings According to Mortality Mortality
Wells Orta Yüksek Significant parenchymal findings Malignancy COPD sPESI
Nonsurvivor (n = 8)
Survivor (n = 59)
n (%)
n (%)
P*
Odds Ratio
7 (87.5) 1 (12.5) 7 (87.5) 1 (12.5) 2 (25.0) 7 (87.5)
52 (88.1) 7 (11.9) 17 (28.8) 4 (6.8) 4 (6.8) 17 (28.8)
1.000
1.061
0.002† 0.482 0.147 0.002†
17.92 1.964 4.583 17.92
95% CI 0.11–9.95 1.97–151.42 0.19–20.15 0.69–30.49 1.97–151.42
*Fisher exact test. †P < 0.01. sPESI indicates Simplified Pulmonary Embolism Severity Index.
168
www.jcat.org
© 2015 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
J Comput Assist Tomogr • Volume 39, Number 2, March/April 2015
Nonsevere Acute Pulmonary Embolism
TABLE 2. Cardiovascular CT Parameters According to Mortality Mortality Nonsurvivor (n = 8)
Qanadli, % PA diameter, mm SVC diameter, mm Azygos diameter, mm RV/LV axial IVC Total diameter, mm
Survivor (n = 59)
Mean ± SD
Median (25%–75% Percentiles)
Mean ± SD
Median (25%–75% Percentiles)
P
Odds Ratio
95% CI
40.94 ± 20.65 30.50 ± 3.58 23.37 ± 4.13 11.00 ± 3.16 1.06 ± 0.31 59.00 ± 4.82
37.50 (22.50–53.75) 29.50 (27.25–33.50) 23.50 (19.5–27.50) 11 (9.25–12.50) 1.14 (0.79–1.31) 59 (56.23–62.75)
44.41 ± 17.94 28.29 ± 8.15 22.22 ± 4.02 10.27 ± 2.41 1.09 ± 0.30 58.17 ± 9.07
50 (35–52.50) 29 (26–32) 22 (19–25) 10 (9.00–12.00) 1.04 (0.86–1.28) 58 (53–62)
0.486 0.456 0.497 0.514 0.900 0.394
0.989 1.041 1.074 1.127 0.734 1.011
0.95–1.03 0.94–1.15 0.89–1.29 0.83–1.51 0.06–9.34 0.93–1.09
Mann-Whitney U test.
significant relationship between the RV/LV axial diameter ratio and the risk of death, which is consistent with another study that reported that an RV/LV diameter ratio of more than 1 was predictive of death only when the embolic burden is low (
Nonsevere Acute Pulmonary Embolism: Prognostic CT Pulmonary Angiography Findings Mehmet Mahir Atasoy, MD,* Nesrin Sariman, MD,† Ender Levent, MD,† Rahmi Çubuk, MD,* Ömer Çelik, MD,‡ Attila Saygi, MD,† Işil Atasoy, MD,§ and Sinan Şahin, MD║ Purpose: To retrospectively evaluate the prognostic parameters of computed tomography (CT) pulmonary angiographic findings in nonsevere (hemodynamically stable) pulmonary embolism (PE) patients and to assess the predictive value of these parameters for mortality within 1 month of the initial diagnosis. Materials and Methods: Retrospectively, 67 consecutive patients (28 men, 39 women; mean age, 63.25 ± 18 years) from 2 centers with nonsevere PE diagnosed using CT and a clinical evaluation were included in the current study. Using consensus reading, 2 readers blinded to the patients' clinical outcomes quantified the right ventricle short axis to left ventricle short axis ratio in the axial plane, vascular measurements, reflux of contrast medium into the inferior vena cava and azygos vein, ventricular septal bowing, and clot load using the Qanadli scoring system. The Simplified Pulmonary Embolism Severity Index (sPESI) and pulmonary parenchymal findings were also evaluated. All CT pulmonary angiographic parameters were compared with the risk of death within 1 month using logistic regression analysis. Results: Fifty-nine patients survived (88.1%), and 8 patients (11.9%) died because of PE. The sPESI and 2 parenchymal findings (multiple wedgeshaped opacities or consolidation accompanied by a wedge-shaped opacity) were significantly related to mortality. In the univariate analysis, neither the cardiovascular CT parameters nor the clot burden was significant between the survivors and nonsurvivors (P > 0.05). Conclusions: In clinically nonsevere PE patients, the sPESI and significant parenchymal findings were the CT parameters related to 1-month mortality. Key Words: pulmonary embolism, nonsevere, hemodynamically stable, computerized tomography pulmonary angiography, prognostic findings (J Comput Assist Tomogr 2015;39: 166–170)
P
ulmonary embolism (PE) is the third most common acute cardiovascular disease.1,2 The results of large randomized trials of thrombolytics3 have indicated that clinical presentation can be considered the most powerful predictor of death caused by PE. A high mortality rate of 50% to 58% has been found in PE patients presenting with hemodynamic instability, and intensive treatment options such as fibrinolytic treatments or mechanical thrombectomy can be provided to these patients.4,5 Although nonsevere (hemodynamically stable) PE patients have a lower but still considerably high mortality rate (8%–15%), they are generally not given intensive treatments.3,4 The mortality rate for a From the *Departments of Radiology and †Respiratory Disease, Maltepe University School of Medicine; ‡Department of Cardiology, Mehmet Akif Ersoy Thoracic and Cardiovascular Surgery Training and Research Hospital; and §Departments of Cardiology and ║Radiology, Dr. Siyami Ersek Thoracic and Cardiovascular Surgery Training and Research Hospital, Istanbul, Turkey. Received for publication August 14, 2014; accepted November 19, 2014. Reprints: Mehmet Mahir Atasoy, MD, Radiology Department, Medical Faculty, Maltepe University, Feyzullah St No. 39 Postal Code 34844 Maltepe, Istanbul, Turkey (e‐mail: [email protected]). This article is an original research. It is not presented at any other congress. There is no funding received. The authors declare no conflict of interest. Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
166
www.jcat.org
subgroup of nonsevere PE patients may be reduced by providing intensive treatment. As computed tomography pulmonary angiography (CTPA) is now the first-line method for diagnosing PE,6,7 evaluating CTPA findings as possible predictors of mortality may be helpful in determining the appropriate subgroup for intensive treatment. Many studies have investigated the prognostic significance of cardiovascular CT parameters and their relationship with patient outcome, but they have not controlled for clinical presentation variables.8–14 Most previous studies have been retrospective studies using radiological data, and few studies have differentiated between severe and nonsevere PE patients and evaluated CT findings as predictors of mortality.15,16 The current study specifically analyzes nonsevere PE patients using a multidisciplinary approach. The objectives of this study were to evaluate the prognostic parameters of CTPA findings in nonsevere acute PE patients with various criteria and to assess the predictive value of these parameters for mortality within 1 month of the initial diagnosis.
MATERIALS AND METHODS Patients Retrospectively, 72 consecutive patients diagnosed as having PE between October 1, 2008, and December 31, 2012, at 2 centers were initially included in our study. Patients with hemodynamic instability were excluded. Blood pressure was taken on admission in all cases, and only hemodynamically stable patients (those with nonsevere PE) were assessed when completing the clinical probability score forms. Patients with chronic PE, determined by CT interpretation and patient history, were also excluded. Patients diagnosed (based on clinical assessment) as having pneumonia, sepsis, diffuse fibrosis, and hemoptysis (caused by a pulmonary hemorrhage) were also excluded because wedge-shaped opacities and consolidation of these diseases can cause misinterpretations of the parenchymal findings. Of the 72 nonsevere PE patients, 5 were excluded from this study (Fig. 1); the final study group included 67 hemodynamically stable PE patients. The institutional review board of our hospital approved this study, with waiver of informed consent. For each patient, the following information was collected: (a) blood pressure values at admission, (b) clinical probability assessment at admission using the Wells Scoring System, (c) patient outcome (survival or death within a 1-month period), and (d) the presence of comorbidity factors (age, presence of cancer or chronic obstructive pulmonary disease [COPD]). Patients with shock and hypotension were considered hemodynamically unstable (severe PE). “Hypotension” was defined as a systolic blood pressure less than 90 mm Hg or a pressure drop of 40 mm Hg or more for longer than 15 minutes not caused by new-onset arrhythmia, hypovolemia, or sepsis.17 Patients were accepted as nonsevere if they did not present with shock or marked hypotension. J Comput Assist Tomogr • Volume 39, Number 2, March/April 2015
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
J Comput Assist Tomogr • Volume 39, Number 2, March/April 2015
Nonsevere Acute Pulmonary Embolism
Cardiovascular CTPA Findings and Clot Burden
FIGURE 1. Flow chart of the non-severe pulmonary embolism patients.
Simplified Pulmonary Embolism Severity Index and Comorbidities Charts were reviewed for the age of the patients and the presence of either malignancy or COPD as a comorbidity. Patient outcome and mortality were retrospectively analyzed on the basis of medical charts and by using the Social Security System to establish whether patients were alive or had died during the follow-up period. The Simplified Pulmonary Embolism Severity Index (sPESI), which is a validated index for estimating 1-month mortality in patients with acute PE, was also calculated for all patients. The test is composed of the following 6 variables, which were weighted equally (1 point per variable): age older than 80 years, history of chronic cardiopulmonary disease, history of cancer, heart rate of 110 or more beats per minute, systolic blood pressure of less than 100 mm Hg, and arterial oxyhemoglobin saturation level of less than 90%. Patients with an sPESI of 1 or higher were classified as high risk and reportedly have a 1-month mortality rate of 10.9%.18
CTPA Technique and Interpretation Twenty-seven CTPA examinations were performed using a 16-detector row CT scanner (Siemens Sensation 16; Siemens AG, Medical Solutions, Erlangen, Germany), and 40 CTexaminations were performed using a 64-MDCT scanner (Aquillion 64; Toshiba Japan); images were acquired in a caudocranial direction. Patients were scanned at 120 kVp and 150 to 250 mA, according to subject body weight, with 1-mm thick sections. The images were reconstructed with a 0.8/1-mm reconstruction interval using a standard reconstruction algorithm. For intravenous access, the antecubital vein and an 18- or 20-gauge catheter were preferred. A bolus of 100 mL of the iodine contrast agent iobitridol (Xenetix 350, 350 mg/mL; Guerbet, Villepinte, France) was injected intravenously via an antecubital vein at a flow rate of 4 mL/s using a dual-head power injector, followed by a 25-mL saline chaser (5 mL/s). Bolus timing was used to optimize pulmonary artery (PA) opacification for all patients. Two radiologists with 8 and 12 years of experience in thoracic radiology interpreted the CT images. Data were transmitted to Workstation Vitrea 2 software (Vital Images Inc, Plymouth, Minn), and the images were reconstructed using postprocessing methods. Both radiologists were blinded to the clinical outcome and evaluated the CT findings by consensus reading. The CT parameters included the right ventricle (RV) short axis–to–left ventricle (LV) short axis (RV/LV) ratio in the axial plane, main PA diameter, azygos vein diameter, superior vena cava (SVC) diameter, and reflux of contrast medium into the inferior vena cava (IVC). The presence of ventricular septal bowing (VSB), pulmonary parenchymal findings, and PA clot load scores were also recorded.
The ventricular diameter measurements were obtained by measuring the short axes of the RV and LV in the axial plane at their widest points. An RV/LV axial diameter ratio cutoff of 1 was used, as recommended in the literature.19,20 We measured the transverse diameter of the main PA (through the midpoint of the pulmonary trunk) and the transverse diameter of the SVC at the level of the azygos vein. We also measured the azygos vein at its widest segment and the IVC just superior to the level of the intrahepatic vein. The IVC diameter was only measured using ultrasonography in previous studies.21,22 Because the IVC rarely appears round on CT images, we calculated the total IVC diameter as the sum of the longest diameter and the short diameter perpendicular to the longest. The PE clot load index was calculated using the method described by Qanadli et al.23 An index higher than 40% was considered to indicate a high embolic burden, which is consistent with previous literature.16 We evaluated the IVC for the presence of reflux. We accepted the VSB as positive when there was both septal flattening and septum deviation convex toward the left ventricle; VSB was considered negative in cases of septum deviation convex toward the right ventricle (normal position).
Pulmonary CT Findings The parenchymal findings were evaluated to check for wedge-shaped opacities (likely pulmonary infarction) and consolidation. A previous report suggested that these 2 parenchymal findings were significantly related to PE.24 Patients with both a wedge-shaped opacity and consolidation and patients with wedge-shaped opacities in multiple locations were considered to have significant parenchymal findings.
Statistical Analysis The NCSS (Number Cruncher Statistical System) 2007 and PASS (Power Analysis and Sample Size) 2008 Statistical Software (Kaysville, Utah) programs were used for statistical analysis. Descriptive statistics (mean, SD, median, frequency, and ratio) were calculated for the patient age, clot load scores, and all CT parameters. Student t test was used to analyze normal distributions, and the Mann-Whitney U test was used for non-normal distributions. Fisher exact test was used for the nominal variables. Pearson correlation analyses were used to determine the relationships between the parameters. The Enter method of logistic regression was used for multivariate analysis. For all tests, a value of P < 0.05 was considered statistically significant.
RESULTS Patients and Clinical Findings All 67 hemodynamically stable patients (28 men, 39 women; mean age, 63.25 ± 18 years) had blood pressures at admission higher than shock levels (range between 170/110 and 90/60). Clinical probability scores at admission were recorded according to the Wells scoring system. The Wells scores varied between 2.5 and 8, and the mean was 5.15 ± 1.59. There were no significant differences in mortality with respect to the clinical probability scores (P > 0.05) according to Fisher exact test. All patients without severe renal dysfunction were initially given low-molecular-weight heparin at weight-adjusted doses, and they continued to receive anticoagulant treatment. Eight patients were referred to the intensive care unit after admission to the hospital. Intravenous heparin was administered to 32 patients (89%). Systemic thrombolysis was performed in only 1 patient. In addition, an IVC filter was inserted in 1 patient (3%). None of the patients underwent mechanical thrombolysis.
© 2015 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
www.jcat.org
167
J Comput Assist Tomogr • Volume 39, Number 2, March/April 2015
Atasoy et al
Nine patients (13.2%) died during the 1-month follow-up period. Death was caused by malignancy in only 1 case and by PE in the other 8 patients (11.8%) according to the independent adjudication committee. The 8 patients who died as a result of PE were considered the nonsurvivor group for the statistical analysis. The average time of death was 14 days after CTPA examination (range, 2–29 days; median, 15 days). With respect to age, 18 patients were older than 70 years (26.8%), and 7 were older than 80 years (10.4%). Five patients were diagnosed as having malignancies (7.4%), and 4 patients were diagnosed as having COPD (9.0%). All patients were diagnosed as having nonsevere PE. Only 2 patients had a heart rate of 110 or more beats per minute, and 3 had a systolic blood pressure less than 100 mm Hg. Three patients had arterial oxyhemoglobin saturation levels less than 90%. There were 24 patients with sPESI of 1 or higher, with 7 belonging to the nonsurvivor group. There were no significant differences between any of the comorbidity factors and the sPESI (Fisher exact test; Table 1).
Table 1. In patients with clinically nonsevere PE, the only CT parameter that predicted mortality was the presence of significant parenchymal findings. Neither the PA CTOI nor any of the other cardiovascular parameters was correlated with the presence of significant parenchymal findings.
DISCUSSION The results of the current study demonstrate that neither cardiovascular CT parameters nor clot load scores are associated with death within 1 month in nonsevere PE patients. However, there is a strong association between the sPESI and significant parenchymal findings, which are known to be somewhat specific to PE. A number of studies evaluating CTPA have attempted to stratify patients without considering clinical presentation.8,9,13,16,25,26 Only one selected a group of patients with severe PE,15 and another evaluated patients with clinically nonsevere PE.16 In the current study, the presence of both a wedge-shaped opacity and consolidation or of multiple wedge-shaped opacities is a predictor of mortality in nonsevere PE patients. In addition, the sPESI was associated with death in nonsevere PE patients within the first month after diagnosis; this finding is consistent with previously conducted nonselective studies. The first month is a critical time for PE patients, and it is the common follow-up period. In addition, 81% of patients have been reported to show complete resolution of PE on CTPA images within 28 days of diagnosis.27 To date, many studies have reported different predictors of mortality, but the actual relationship of these characteristics with mortality is debated in the literature. Arterial blood gas analysis effectively predicts the prognosis for acute PE.28 Although parenchymal disease provides information on hemodynamic tolerance to RV dysfunction, it was not previously studied as a cause of mortality. In the current study, we did not evaluate the nonspecific findings of pulmonary parenchyma; instead, we evaluated only wedge-shaped opacities and consolidation. These 2 parameters are somewhat specific findings to PE, particularly in the absence of pneumonia, sepsis, diffuse fibrosis, and hemoptysis (caused by a pulmonary hemorrhage).23 Because we only included patients with clinically nonsevere PE in the present study, a comparison of our results with the majority of similar studies in the literature (in which the clinical status at admission is not considered) is difficult. The current study used the axial RV/LV diameter ratio, which is reportedly no less accurate than the reformatted 4-chamber RV/LV diameter ratio for predicting 1-month mortality after PE.29 We did not find a
Cardiovascular CT Measurements and Clot Burden The Qanadli scores were calculated to be between 3% and 95% with a mean of 43.99% ± 18.14%. Survivors had a mean Pulmonary Artery CT Obstruction Index (PA CTOI) of 44.41%, and nonsurvivors had a mean PA CTOI of 40.94% (P > 0.05). There was no significant difference in the mortality rate between high and low embolic burden patients (P = 0.44; Mann-Whitney U test). The RV/LV ratio in the axial plane, main PA diameter, azygos vein diameter, and IVC and SVC diameters were evaluated in all patients. There were no significant differences between the survivors and nonsurvivors with respect to these factors (MannWhitney U test). There were no significant differences in the RV/LV ratios according to the cutoff value and mortality as determined by Fisher exact test (P = 1.000; P > 0.05 for both). The cardiovascular CT parameters are listed in Table 2, along with the other quantitative measurements. Twenty-one patients (35.6%) had IVC reflux. However, this reflux was not predictive of mortality. Thirty patients (50.8%) showed VSB. We interpreted a smooth interventricular septum as VSB; the presence of VSB was not related to mortality in clinically nonsevere PE patients.
Pulmonary Parenchymal CT Findings Whereas 7 patients (7/8, 87.5%) in the nonsurvivor group had significant parenchymal findings, only 17 patients (17/59, 28.8%) in the survivor group had significant parenchymal findings (P = 0.002) (Fig. 2). The odds ratio was 17.92, as shown in
TABLE 1. Comorbidity Factors and Parenchymal Findings According to Mortality Mortality
Wells Orta Yüksek Significant parenchymal findings Malignancy COPD sPESI
Nonsurvivor (n = 8)
Survivor (n = 59)
n (%)
n (%)
P*
Odds Ratio
7 (87.5) 1 (12.5) 7 (87.5) 1 (12.5) 2 (25.0) 7 (87.5)
52 (88.1) 7 (11.9) 17 (28.8) 4 (6.8) 4 (6.8) 17 (28.8)
1.000
1.061
0.002† 0.482 0.147 0.002†
17.92 1.964 4.583 17.92
95% CI 0.11–9.95 1.97–151.42 0.19–20.15 0.69–30.49 1.97–151.42
*Fisher exact test. †P < 0.01. sPESI indicates Simplified Pulmonary Embolism Severity Index.
168
www.jcat.org
© 2015 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
J Comput Assist Tomogr • Volume 39, Number 2, March/April 2015
Nonsevere Acute Pulmonary Embolism
TABLE 2. Cardiovascular CT Parameters According to Mortality Mortality Nonsurvivor (n = 8)
Qanadli, % PA diameter, mm SVC diameter, mm Azygos diameter, mm RV/LV axial IVC Total diameter, mm
Survivor (n = 59)
Mean ± SD
Median (25%–75% Percentiles)
Mean ± SD
Median (25%–75% Percentiles)
P
Odds Ratio
95% CI
40.94 ± 20.65 30.50 ± 3.58 23.37 ± 4.13 11.00 ± 3.16 1.06 ± 0.31 59.00 ± 4.82
37.50 (22.50–53.75) 29.50 (27.25–33.50) 23.50 (19.5–27.50) 11 (9.25–12.50) 1.14 (0.79–1.31) 59 (56.23–62.75)
44.41 ± 17.94 28.29 ± 8.15 22.22 ± 4.02 10.27 ± 2.41 1.09 ± 0.30 58.17 ± 9.07
50 (35–52.50) 29 (26–32) 22 (19–25) 10 (9.00–12.00) 1.04 (0.86–1.28) 58 (53–62)
0.486 0.456 0.497 0.514 0.900 0.394
0.989 1.041 1.074 1.127 0.734 1.011
0.95–1.03 0.94–1.15 0.89–1.29 0.83–1.51 0.06–9.34 0.93–1.09
Mann-Whitney U test.
significant relationship between the RV/LV axial diameter ratio and the risk of death, which is consistent with another study that reported that an RV/LV diameter ratio of more than 1 was predictive of death only when the embolic burden is low (
