Markers of right ventricular dysfunction predict 30-day adverse prognosis of pulmonary embolism on pulmonary computed tomographic angiography

Abstract

To investigate the value of parameters of the pulmonary artery and right ventricular function in predicting the 30-day poor prognosis of patients with acute pulmonary embolism (APE). The heart rate, respiratory rate, systolic blood pressure, Wells score for APE, history of recent operation or immobilization, history of cancer, respiratory failure, smoking were significantly (P < .05) different among the control, good prognosis, and poor prognosis groups. The maximal short diameter of the right and left ventricle (RVD/LVD) ratio (P < .001) and left pulmonary artery (LPA) (P = .01) were significantly different between the good and poor prognosis groups. Systolic blood pressure (odds ratio [OR]: 0.98, P = .045) and the RVD/LVD ratio (OR: 12.57, P = .02) were significant independent risk factors for poor prognosis. The risk for poor prognosis significantly increased when the RVD/LVD ratio was >1.11 (cutoff value) with the area under the curve (AUC) of 0.71 (95% confidence interval [CI]: 0.61–0.80, P < .001). LPA (OR: 9.12, P = .01) and RVD/LVD (OR: 4.62, P = .012) were the significant independent risk factors for poor prognosis in the central pulmonary embolism. The LPA of 2.1 cm had the highest predictive value for poor prognosis in the central APE (AUC: 0.68; sensitivity 84.6%; specificity 53.1%). The RVD/LVD ratio and systolic blood pressure are significant risk factors for short-term prognosis in patients with APE. When the LPA is >2.1 cm in the central APE or the RVD/LVD is >1.11, the risk of poor prognosis increases, which can be used as important indicators for predicting the prognosis of patients with APE. Two hundred forty-three APE patients and 61 patients without APE who underwent computed tomographic pulmonary angiography (CTPA) were retrospectively enrolled as the experimental and the control group, respectively. APE patients who were followed up at the 30-day time point were divided into the good prognosis (n = 195) and poor prognosis group (n = 32). The main pulmonary artery (MPA) to the aorta (AO) ratio, maximal diameter of the LPA and right pulmonary artery (RPA), ratio of the RVD/LVD and the height and volume of the pulmonary artery (PAh and PAV, respectively) were analyzed after indexing to the body surface area.

1. Introduction

Acute pulmonary embolism (APE) is the third most common cause of cardiovascular death following the coronary artery disease and stroke, necessitating emergency treatment.^[1–3] Pulmonary arterial hypertension and right ventricular dysfunction caused by the pulmonary embolism are the main causes of poor prognosis in patients with severe pulmonary embolism.^[4] At present, the computed tomographic pulmonary angiography (CTPA) is the gold-standard method commonly used in evaluating pulmonary embolism. It is vital to use CTPA to diagnose, perform risk stratification, and evaluate adverse prognosis in patients with pulmonary embolism.^[5] Recently, CTPA has been considered as an efficient tool for the diagnosis of pulmonary hypertension through volume analysis of the main pulmonary artery (MPA).^[6] The presence of elevated pulmonary pressures in chronic lung or cardiac disease has also been associated with an increased risk of mortality.^[7] The pulmonary artery obstruction index (PAOI) is the CTPA occlusion index which quantifies the score of blood clot in evaluating the degree of pulmonary embolism obstruction. The ratio of the right ventricular diameter to the left ventricular diameter (RVD/LVD) as an index to evaluate the right ventricular function is simpler and faster than the PAOI parameter, however, the correlation of the above 2 indexes with adverse prognosis is not uniform.^[2,3,8–12] Apfaltrer et al^[8] believed that the obstruction index was not associated with a poor clinical prognosis. Coutance et al^[10] found that the diameter ratio of RVD/LVD ≥ 1 was not associated with mortality, whereas Hefeda et al^[11] and Kang et al^[12] concluded that the RVD/LVD ratio and PAOI could predict short-term mortality. These controversial beliefs suggested the necessity of further investigation on the clinical value of these parameters.

In addition to the RVD/LVD ratio and the width of pulmonary artery trunk, the parameters related to the right ventricular function include the inferior vena cava contrast agent reflux, the height and volume of the pulmonary artery (PAV) trunk, and the shape and filling of the pulmonary vein. There is a lack of research in the literature on whether these parameters are related to poor prognoses of patients with pulmonary embolism. It was hypothesized that all these parameters were associated with the 30-day poor prognosis of patients with APE. This study was consequently conducted to investigate the clinical significance of these parameters related to the right ventricular function of patients with APE using the 256-slice spiral CTPA.

2. Materials and methods

2.1. Subjects

This retrospective case-control 1-center study was performed between May 2018 and May 2020 with approval of the ethics committee of the Second Hospital of Hebei Medical University (2022-R732), and all patients or their family members had signed the informed consent to participate. The inclusion criteria for the APE group were patients with APE in line with the 2019 European Heart Association Guidelines for Diagnosis and Treatment of Acute Pulmonary Embolism,^[2,13] who had undergone CTPA examination, and with detailed clinical and follow-up data. The exclusion criteria were patients with other cardiac diseases causing cardiac enlargement, including chronic pulmonary heart disease, rheumatic heart disease, congenital heart disease, and cardiomyopathy (hypertrophic cardiomyopathy, dilated cardiomyopathy, restrictive cardiomyopathy, etc) and previous heart failure history. Patients with poor-quality CTPA imaging and a history of chronic pulmonary embolism were also excluded. The control group was composed of patients without APE who had undergone CTPA examination during the same period because of chest pain or D-dimer elevation and with no pulmonary embolism, and the exclusion criteria were the same as those for the APE group. Poor prognosis was defined as presence of at least one of the following events within 30 days^[14,15]: death, cardiopulmonary resuscitation, endotracheal intubation, demand for vasopressors (more than 5 μG/kg) in patients with systemic hypotension, and reperfusion therapy.

APE was defined as presence of an endoluminal, low-attenuation filling defect with partial or total obstruction of a pulmonary artery. The patient usually had an acute onset or a history of lying in bed after surgical operation. CT scan showed occlusion of a pulmonary artery, a track sign formed by the embolus and the wall of the peripheral pulmonary artery for the peripheral type of pulmonary embolism, and thickening of the corresponding pulmonary artery and its branches. The embolus in the acute obstructive lumen was completely wrapped by the contrast agent, and the embolus in the arterial lumen formed an obtuse angle with the arterial wall.^[2,16–18] Chronic pulmonary embolism was diagnosed if, after more than 3 months of standardized anticoagulation treatment, medical imaging had confirmed the existence of chronic thrombosis, the average pulmonary artery pressure (mPAP) measured using the right heart catheterization was ≥25 mm Hg, and other diseases had been excluded such as vasculitis and pulmonary artery sarcoma.^[2,18] CT scanning showed that the lumen of the pulmonary artery containing a embolus was narrowed or truncated compared with that of the uninvolved pulmonary artery. The angle between the embolus and the wall was obtuse, a filling defect was wrapped by the contrast agent, and a linear filling defect in the lumen was visible.^[17]

2.2. Examination methods

Computed tomography angiography scan was performed with the Philips 256i CT scanner (Brilliance iCT, Philips Healthcare, Cleveland, OH) with the patient in the supine position. The scanning range was from the thoracic entrance to the angle of the costal diaphragm. The nonionic contrast agent iodohydrin was injected with a high-pressure syringe at a speed of about 4 to 5 mL/s with a dosage of 50ml (adjusted according to the actual weight of the patient). CT scanning was conducted at the end of a single exhalation, with the breath holding time of 4 to 7 seconds. Non-cardiac gating and automatic triggering scanning technology was used. The scanning parameters were tube current 200 to 300 mAs/revolution, tube voltage 120 kv, collimation 128 × 625, pitch 0.16 to 0.2, rotation time 270 to 330 ms, matrix 512 × 512, and a display field 350 mm.

2.3. Imaging processing and measurement

The Philips EBW 4.5 working station was used for imaging processing and measurement of the pulmonary artery and right cardiac parameters (Figs. 1 and 2). All images were evaluated by 2 experienced radiologists independently, and the average values were calculated for outcome analysis.

Figure 1.

Data measurement was performed on the axial plane images of computed tomography pulmonary angiography (CTPA). (A–C) The maximal diameter of the main pulmonary artery (MPA) and the aorta (AO) (A), right and left pulmonary arteries (RPA and LPA) (B and C) were measured. The diameter of MPA is 25.0 mm, and the AO diameter is 28.3 mm (A). The maximal diameter of RPA is 19.6 mm (B), and the maximal diameter of LPA is 19.0 mm (C). (D) The maximal right ventricular diameter (RVD) and left ventricular diameter (LVD) were measured. The width of right ventricle is 46.7 mm, and the width of left ventricle is 32.9 mm. The ratio of RVD/LVD is 1.42. RVD/LVD = maximal short diameter of the right and left ventricles.

Figure 2.

Markers of right ventricular dysfunction was performed on the images of computed tomography pulmonary angiography (CTPA). (A) The distance from the pulmonary valve to the main pulmonary branch is 52.1 mm. (B) Pulmonary artery volume was measured on 3D imaging including the sum of the volumes of the main, right and left pulmonary arteries. (C) The ventricular septum was bowing (arrow) towards the left ventricle. (D) Reflux of contrast medium occurred into the inferior vena cava and the distal hepatic vein.

All of the following parameters were measured on the axial CT imaging. The MPA diameter, the left pulmonary artery diameter (LPA), and the right pulmonary artery diameter (RPA) were determined from the maximal cross-sectional areas on the imaging. The diameter of the ascending aorta (AO) was measured on the same axial imaging. The maximal diameters of RV and LV were derived separately from the maximal cross-sectional areas of RV and LV.^[14]

The height of the pulmonary artery (PAh) trunk was measured on the multi-planar reconstruction (MPR) images, which indicated the distance from the pulmonary valve to the bifurcation of the left and right pulmonary arteries. The PAV was the sum of the volumes of the main, right and left pulmonary arteries.^[6,19] On 3D image of the pulmonary artery, the pulmonary valve, and the branch-off point of LPA and RPA was used as the border to calculate the pulmonary artery volume. The morphology of the interventricular septum was observed (Fig. 3). There were 3 types of ventricular septal morphology: the mild right ventricular process which was the normal type, straight ventricular process and left ventricular process. The inferior vena cava reflux was determined on axial plane images when the contrast agent flowed back into the inferior vena cava.

Figure 3.

Flowchart of patient enrollment.

Pulmonary embolism was divided into the central and peripheral type based on the location of the embolism. According to the classification of embolic vessels, the central type was defined as the embolism in the MPA, left and RPA and lobar pulmonary artery, whereas the peripheral type was referred to as the embolism in the pulmonary artery below the lobar grade.^[20] The PAOI was measured by using the method stated by Qanali et al,^[21] and bilateral pulmonary arteries were divided into 3 upper lobes, 2 lingulas (or middle lobes) and 5 lower lobes (10 segments in total). The embolism of each segment of the pulmonary artery was rated as 1 point, and the score of the embolism in the proximal pulmonary artery was the sum of pulmonary segments. The degree of pulmonary embolism was based on a 2-grade method (0 for no obstruction, 1 for partial occlusion and 2 for complete occlusion). The calculation formula was [∑(n*d)/40] × 100%, with n as the proximal pulmonary artery embolism and d as the degree of arterial occlusion.^[21]

The Wells score^[22] was used to evaluate the diagnostic efficacy of APE, and the criteria of the score are: a previous history of pulmonary embolism or deep venous thrombosis, heart rate ≥ 100 beats/min, history of operation or braking in the past 4 weeks, hemoptysis, active stage of malignant tumor, deep-venous-thrombosis related symptoms, and low possibility of diagnosis other than the pulmonary embolism. Each item was assigned 1 point.

2.4. Statistical analysis

The statistical analysis was performed with the SPSS 25.0 (IBM, Chicago, IL). Normal distribution of the variables was tested with the “exploration” function of the SPSS software. Measurement data were presented as the mean ± standard deviation if in the normal distribution and tested with the independent t test or ANOVA. If not in the normal distribution, the measurement data were presented as the median and interquartile range and tested with the Kruskal–Wallis H test. The variables were compared between 2 groups only when the P value for the ANOVA/Kruskal–Wallis H test was significant. The Bonferroni correction was used to adjust the significance values for multiple tests. Categorical variables were presented as frequency and percentages and tested with the Chi square test. The interclass correlation coefficient was used to test the repeatability of the measurement. Adverse prognosis of patients with APE was treated as dependent variables. First, the univariate logistic regression analysis was carried out on factors associated with a poor prognosis, and then, the multivariate logistic regression analysis was carried out on the above significant indicators. The receiver operator characteristics (ROC) curve analysis was performed to analyze the risk factors and cutoff values for poor prognosis of patients with APE. The optimal cutoff value was chosen as the one that achieved the maximal Youden index. The significant P value was set at <.05.

3. Results

3.1. Participant characteristics

One hundred ninety-five (85.9%) patients with APE had a good 30-day prognosis and were assigned to the good prognosis group, including 96 (49.2%) female and 99 (50.8%) male patients aged 17–90 (mean 64) years (Fig. 3). Thirty-two (14.1%) patients with APE experienced a poor 30-day prognosis and were assigned to the poor prognosis group, including 20 (62.5%) female and 12 (37.5%) male patients with an age range 38 to 85 (mean 68) years. In the control group, 61 patients without APE were enrolled, including 28 males (45.9%) and 33 females (54.1%) aged 53 to 72 (mean 65) years (Table 1). The heart rate (P < .001), respiratory rate (P < .001), systolic blood pressure (P = .001), Wells score for APE (P < .001), history of recent operation or immobilization (P = .001), history of cancer (P = .02), respiratory failure (P < .001), and smoking (P = .02) among the 3 groups were significantly different. Between the good and the poor prognosis groups, the systolic blood pressure and embolism types (P = .001) were significantly different, with the poor prognosis group having significantly lower systolic pressures (118.25 ± 23.68 vs 133.09 ± 18.65 mm Hg, P = .001) and significantly (P = .001) more central embolism type but less peripheral type (Table 1). In the good prognosis group, 98 (50.26%) patients presented with the central pulmonary embolism, whereas in the poor prognosis group, 26 (81.25%) patients presented with the central pulmonary embolism.

Table 1.

Comparison of clinical data among the control, good prognosis and poor prognosis groups.

		Control (n = 61)	Good prognosis (n = 195)	Poor prognosis (n = 32)	P
Age (yr, mean)		65 (19.00)	65 (11.00)	68 (15.00)	.16
Sex	F (n, %)	33 (54.10%)	96 (49.20%)	20 (62.50%)	.35
	M (n, %)	28 (45.90%)	99 (50.80%)	12 (37.50%)
Clinical features	Shortness of breath/dyspnea	38 (62.30%)	132 (68.80%)	21 (65.60%)	.64
	Hemoptysis	1 (1.64%)	11 (5.60%)	2 (6.30%)	.42
	Amaurosis/Syncope	5 (8.20%)	23 (11.80%)	7 (21.90%)	.15
	Palpitation	10 (16.39%)	23 (11.80%)	5 (15.60%)	.59
	Chest pain	5 (8.20%)	32 (16.40%)	4 (12.50%)	.27
	Lower limb edema/pain	4 (6.56%)	25 (12.80%)	2 (6.30%)	.26
	Heart rate (times/min)	78 (20.00)	83 (16.00)*	90 (20.00)*#	＜.001
	Respiratory rate (times/min)	20 (1.00)	20 (2.00)*	20 (2.75)*	＜.001
	Systolic blood pressure (mm Hg)	132.57 ± 21.55	133.09 ± 18.65	118.25 ± 23.68#	.001
	Wells score	0 (1.00)	1 (2.00)*	2 (1.00)*#	＜.001
Risk factors	History of pulmonary embolism or deep venous thrombosis	0	13 (6.70%)*	1 (3.10%)	.10
	Recent operation or immobilization	4 (6.56%)	30 (15.40%)	12 (37.50%)*#	.001
	Cancer history	0	17 (8.70%)*	5 (15.60%)*	.02
	Respiratory and circulatory failure	1 (1.64%)	36 (18.50%)*	10 (31.30%)*	＜.001
	Heart failure	11 (18.03%)	27 (13.80%)	4 (12.50%)	.68
	Diabetes	10 (16.39%)	29 (14.90%)	7 (21.90%)	.60
	Transient ischemic attack	39 (63.93%)	32 (16.40%)	5 (15.60%)	.45
	Hypertension	31 (50.82%)	100 (51.30%)	12 (37.50%)	.35
	Peripheral vascular disease	6 (9.84%)	27 (13.80%)	1 (3.10%)	.19
	Smoking history	9 (14.75%)	57 (29.20%)*	4 (12.50%)#	.02
Types	Central	-	98 (50.26%)	26 (81.25%)	.001
	Peripheral	-	97 (49.74%)	6 (18.75%)

P, comparison among 3 groups;

^*P < .05 compared with the control group;

^#P < .05 compared with patients with a good prognosis. Data presentation: frequency (%). When the data were not in the normal distribution, they were presented as a median (interquartile range). Heart failure here indicates the newly appeared heart failure which occurred after admission.

Two patients were treated with conservative physical therapy due to anticoagulation contraindication, 211 patients received the anticoagulant therapy, and 14 patients were treated with intravenous thrombolysis + anticoagulants.

3.2. Morphological and functional parameters of the RV and pulmonary artery

After measurement of the relevant parameters, the repeatability test showed good consistence between the 2 radiologists, with the lowest interclass correlation coefficient value of 0.87 (range 0.84–0.90) (Table 2). The MPA/AO ratio, AO, MPA, RPA, LPA, pulmonary artery volume (PAV), RVD/LVD ratio, and ventricular septal displacement were significantly different among 3 groups (P < .001, .04, <.001, <.001, <.001, <.001, <.001, respectively) (Table 3).

Table 2.

ICC for parameters of the right ventricle and pulmonary artery.

Variables	ICC	95%CI	P
AO (cm/m²)	0.97	0.96–0.98	＜.001
MPA (cm/m²)	0.93	0.92–0.95	＜.001
RPA (cm/m²)	0.93	0.91–0.95	＜.001
LPA (cm/m²)	0.87	0.84–0.90	＜.001
PAh (cm/m²)	0.88	0.85–0.91	＜.001
PAV (mL/m²)	0.97	0.97–0.98	＜.001
RVD/LVD	0.89	0.86–0.92	＜.001

AO = aorta, CI = confidence interval, ICC = interclass correlation coefficient, LPA = left pulmonary artery, LVD = left ventricular maximal short diameter, MPA = main pulmonary artery, PAh = pulmonary artery height, PAV = pulmonary artery volume, RPA = right pulmonary artery, RVD = right ventricular maximal short diameter, RVD/LVD = maximal short diameter of the right and left ventricles.

Table 3.

Parameters of the right ventricle and pulmonary artery among the control, good and poor prognosis groups (mean ± standard deviation).

Parameters	Control (n = 61)	APE (n = 227)		P
		Good prognosis (n = 195)	Poor prognosis (n = 32)
MPA/AO	0.75 (0.14)	0.82 (0.19)*	0.84 (0.16)*	＜.001
AO (cm)	3.42 ± 0.46	3.57 ± 0.35*	3.41 ± 0.42	.04
MPA (cm)	2.48 (0.54)	2.92 (0.67)*	2.98 (0.68)*	＜.001
RPA (cm)	2.04 ± 0.30	2.28 ± 0.35*	2.35 ± 0.30*	＜.001
LPA (cm)	1.97 ± 0.25	2.08 ± 0.29*	2.24 ± 0.41*#	＜.001
PAh (cm)	5.56（0.83）	5.66（0.81）	5.76（0.91）	.46
PAV (mL)	57.30 (23.75)	68.40 (24.90)*	66.20 (28.25)*	＜.001
RVD/LVD	1.06 (0.33)	1.10 (0.39)*	1.36 (0.61)*#	＜.001
PAOI	-	0.28 (0.43)	0.46 (0.27)	.007
VSD	22 (36.07%)	131 (67.18%)*	25 (78.13%)*	＜.001
VCR	41 (67.21%)	101 (51.79%)	18 (56.25%)	.11

AO = aorta, APE = acute pulmonary embolism, LPA = left pulmonary artery, LVD = left ventricular maximal short diameter, MPA = main pulmonary artery, PAh = pulmonary artery height, PAOI = pulmonary artery occlusion index, PAV = pulmonary artery volume, RPA = right pulmonary artery, RVD = right ventricular maximal short diameter, RVD/LVD = maximal short diameter of the right and left ventricles, VSD = ventricular septal displacement, VCR = Vena cava reflux. P, comparison among 3 groups;

^*P < .05 compared with the control group;

^#P < .05 compared with patients with a good prognosis. Data were presented as mean ± standard deviation when they are in the normal distribution and as a median (interquartile range) when they are not in the normal distribution.

After indexing to the body surface area, the MPA, MPA/AO ratio, RPA, LPA, PAV, and RVD/LVD were significantly different among 3 groups (P < .001, .001, <.001, .003, <.001, <.001, respectively) (Table 4). Compared with the good prognosis group, the RVD/LVD ratio (P < .001) and LPA (P = .01) were significantly higher in the poor prognosis group (Table 4).

Table 4.

Parameters of pulmonary artery and right ventricle after indexing to the body surface area.

Variables	Control (n = 61)	APE (n = 227)		P
		Good prognosis (n = 195)	Poor prognosis (n = 32)
MPA/AO	0.43 (0.09)	0.47 (0.11)*	0.49 (0.11)*	.001
AO (cm/m²)	1.98 ± 0.27	2.06 ± 0.33	2.05 ± 0.25	.19
MPA (cm/m²)	1.45 (0.32)	1.66 (0.43)*	1.68 (0.34)*	＜.001
RPA (cm/m²)	1.18 ± 0.18	1.31 ± 0.25*	1.35 ± 0.19*	＜.001
LPA (cm/m²)	1.14 (0.19)	1.17 (0.24)	1.28 (0.25)*#	.003
PAh (cm/m²)	3.21 ± 0.56	3.25 ± 0.50	3.29 ± 0.45	.72
PAV (mL/m²)	31.88 (13.30)	39.02 (14.63)*	40.33 (16.55)*	＜.001
RVD/LVD	0.61 (0.21)	0.63 (0.27)	0.80 (0.42)*#	＜.001

AO = aorta, APE = acute pulmonary embolism, LPA = left pulmonary artery, LVD = left ventricular maximal short diameter, MPA = main pulmonary artery, PAh = pulmonary artery height, PAV = pulmonary artery volume, RPA = right pulmonary artery, RVD = right ventricular maximal short diameter, RVD/LVD = maximal short diameter of the right and left ventricles.

^*P < .05 compared with the control group;

^#P < .05 compared with patients with good prognosis. Data were presented as mean ± standard deviation when they are in the normal distribution and as a median (interquartile range) when they are not in the normal distribution.

3.3. Correlation of different parameters with poor prognosis

The univariate logistic analysis demonstrated that the systolic blood pressure and RVD/LVD were significantly (P < .001) associated with a poor prognosis of APE. The multivariate logistic regression analysis revealed that systolic blood pressure (odds ratio [OR]: 0.98, 95% confidence interval [CI]: 0.96–1.00, P = .045) and RVD/LVD (OR: 12.57, 95% CI: 1.49–106.28, P = .02) were significant independent risk factors for poor prognosis (Table 5).

Table 5.

Logistic regression analysis of factors associated with poor prognosis.

Parameters	Univariate analysis			Multivariate analysis
	B	OR (95%CI)	P	B	OR (95%CI)	P
Heart rate (times/min)	0.04	1.04 (1.01–1.06)	.002	0.02	1.02（0.99–1.05）	.27
Respiratory rate (times/min)	0.07	1.08 (0.91–1.27)	.38
Systolic blood pressure (mm Hg)	−0.04	0.96 (0.94–0.98)	＜.001	−0.02	0.98 (0.96–1.00)	.045
History of APE or deep venous thrombosis	−0.78	0.45 (0.06–3.58)	.45
Recent operation or immobilization	1.19	3.30 (1.46–7.45)	.004	1.07	2.92（0.92–9.29）	.07
Cancer history	0.66	1.94 (0.66–5.69)	.23
Respiratory failure	0.70	2.01 (0.88–4.61)	.10
Smoking	−1.06	0.35 (0.12–1.03)	.06
Types	−1.46	0.23 (0.09–0.59)	.002	−0.87	0.42 (0.11–1.59)	.20
Wells score	0.56	1.75 (1.19–2.58)	.005	0.24	1.27 (0.70–2.30)	.44
PAOI	2.45	11.60 (1.89–71.30)	.008	−0.38	0.69 (0.38–12.24)	.80
VSD	0.56	1.75 (0.72–4.25)	.22
Standardized：
MPA/AO	0.87	2.39 (0.07–82.85)	.63
MPA	0.26	1.29 (0.45–3.75)	.64
RPA	0.62	1.86 (0.40–8.63)	.43
LPA	1.88	6.53 (1.20–35.72)	.03	0.62	1.86 (0.20–17.24)	.59
PAh	0.02	1.02 (0.94–1.10)	.64
PAV	0.00	1.00 (0.98–1.03)	.87
RVD/LVD	3.19	24.33 (4.99–118.76)	＜.001	2.53	12.57 (1.49–106.28)	.02

AO = aorta, APE = acute pulmonary embolism, CI = confidence interval, LVD = left ventricular maximal short diameter, MPA = main pulmonary artery, OR = odds ratio, PAh = pulmonary artery height, PAOI = pulmonary artery occlusion index, PAV = pulmonary artery volume, RPA = right pulmonary artery, VSD = ventricular septal displacement, LPA = left pulmonary artery, RVD = right ventricular maximal short diameter, RVD/LVD = maximal short diameter of the right and left ventricles.

3.4. Predictive value of the RVD/LVD ratio for poor prognosis

The ROC curve analysis of the RVD/LVD ratio revealed that in patients with severe APE, the risk for poor prognosis increased significantly when the RVD/LVD ratio was >1.11 (the cutoff value) with the area under the curve (AUC) of 0.71 (95% CI: 0.61–0.80), a sensitivity of 81.3% and a specificity of 50.3% (Fig. 4), whereas the AUC of the systolic blood pressure was 0.69 (95% CI: 0.59–0.80, P = .001) ，cutoff point was 125.5 mm Hg, with a sensitivity of 71.9% and a specificity of 62.1%.

Figure 4.

The receiver operating characteristics (ROC) cure analysis of the ratio of the maximal short diameter of the right and left ventricles (RVD/LVD) for predicting poor prognosis of patients with acute pulmonary embolism.

3.5. Comparison of parameters between the central and peripheral APE groups

Subgroup analysis of the APE patients with the central and peripheral pulmonary embolism types showed that MPA, MPA/AO, RPA, LPA, PAV, and RVD/LVD were significantly (P = .001, .02, .006, .001, .02 and <.001, respectively) greater in patients with the central than those with the peripheral pulmonary embolism (Table 6). The logistic regression analysis revealed that LPA (OR: 9.12, 95% CI: 1.71–48.52, P = .01) and RVD/LVD (OR: 4.62, 95% CI: 1.40–15.21, P = .012) were the significant independent risk factors for the poor prognosis in patients with the central pulmonary embolism. The ROC curve analysis showed that the AUC was 0.68 for LPA with a sensitivity 84.6%, a specificity 53.1%, and a cutoff value 2.1cm (P = .004). In addition, the AUC was 0.65 for RVD/LVD with a sensitivity 76.9%, a specificity 48%, and a cutoff value 1.17 (P = .016).

Table 6.

Parameters of pulmonary artery and right ventricle after indexing to the body surface area between the central and peripheral types of APE.

Variables	Total (n = 243)	APE		P
		Central (n = 124)	Peripheral (n = 103)
MPA/AO	0.46 (0.11)	0.49 (0.12)	0.45 (0.11)	.02
AO (cm/m²)	2.04 ± 0.31	2.08 ± 0.31	2.03 ± 0.32	.26
MPA (cm/m²)	1.61 (0.41)	1.72 (0.38)	1.56 (0.44)	.001
RPA (cm/m²)	1.29 (0.32)	1.37 (0.31)	1.26 (0.32)	.006
LPA (cm/m²)	1.17 (0.24)	1.22 (0.24)	1.13 (0.26)	.001
PAh (cm/m²)	3.21 (0.63)	3.22 (0.58)	3.19 (0.77)	.69
PAV (mL/m²)	37.64 (15.23)	41.05 (16.65)	37.35 (14.40)	.02
RVD/LVD	0.64 (0.26)	0.76 (0.33)	0.58 (0.23)	＜.001

The data were presented in median (interquartile range) or mean ± standard deviation.

AO = aorta, APE = acute pulmonary embolism, LPA = left pulmonary artery, MPA = main pulmonary artery, PAh = height of the pulmonary artery, PAV = volume of the pulmonary artery, RPA = right pulmonary artery, RVD/LVD = maximal short diameter of the right and left ventricles.

4. Discussion

Currently, CTPA is the most commonly used diagnostic tool for patients with suspected APE.^[23–25] It is advisable to use this imaging method to evaluate the relevant parameters for stratifying the risk of complications in these patients or to administer personalized treatment plan by evaluating the severity of the disease.^[26,27] The right ventricular function is generally considered as a predictor of adverse outcomes.^[28–30] If APE is not treated timely, the pulmonary artery pressure may continue to rise, leading to irreversible remodeling of the pulmonary vessels, which can further aggravate the right heart failure.^[31] Our study found that PAV was a potential indicator for the diagnosis of APE. The RVD/LVD ratio can be used as a valid parameter in daily imaging evaluation of APE. Systolic blood pressure and the RVD/LVD ratio were significant risk factors for poor prognosis of APE patients.

Clinical manifestations of the APE are diverse. Patients may have chest tightness, shortness of breath/dyspnea, hemoptysis, chest pain, syncope or edema of both lower limbs, often accompanied by signs such as accelerated heart rate and breathing, or sudden shock.^[32] Dyspnea was still the primary symptom of APE patients with an incidence of about 67.4%, whereas the typical triad of pulmonary embolism (dyspnea, chest pain, and hemoptysis) was only 1.8% in our study. In fact, hemodynamic status remains one of the most important short-term prognostic factor for patients with acute PE. Quezada confirms that having a low systolic blood pressure (<110 mm Hg) was associated with an increased risk for death (both all-cause and PE-related mortality), but having a high systolic blood pressure (>130 mm Hg) was associated with lower risks.^[33] This study showed that when systolic blood pressure was lower than 125.5 mm Hg, the risk of short-term adverse prognosis was increased. The treatment methods of patients in this study involved physical therapy, anticoagulation, intravenous thrombolysis plus anticoagulation. Clinical research showed that^[34] anticoagulation combined with thrombolytic therapy was better than anticoagulation alone in the treatment of pulmonary embolism. No matter what kind of treatment was used, almost all patients’ symptoms of acute right ventricular pressure overload gradually disappeared, and the hemodynamic benefits of thrombolytic drugs were limited only to the first few days after initiation of the treatment.^[35]

Pulmonary embolism can cause right heart overload, right ventricular dilatation, decreased systolic performance, and subsequent right heart dysfunction which is an important predictor of adverse outcomes of APE. On CT imaging, the RVD/LVD ratio is a well-recognized index to measure RVD and can be used as a surrogate index for RV dysfunction, with the most widely accepted limits as 0.9 and 1.0.^[11,36,37] A meta-analysis found that an increase in the ratio of RVD/LVD measured axially on CTPA is the most powerful independent predictor of all-cause mortality of APE,^[38] however, Coutance et al^[10] found that RVD/LVD ≥ 1 was not associated with mortality. In our study, it was found that the heart rate, recent surgery or immobilization, systolic pressure, types, Wells score, PAOI, LPA and RVD/LVD were important risk factors for poor prognosis of patients with APE, with systolic pressure, and RVD/LVD as 2 independent risk factors for poor prognosis of APE patients. Although Hefeda and Elmasry^[11] proposed that RVD/LVD > 1.2 was closely associated with the 30-day mortality, our study found that when the RVD/LVD ratio was >1.11, the risk of poor prognosis in patients with APE was significantly increased, and this ratio could be used as an effective parameter in daily imaging evaluation of prognosis. In clinical application, the way to obtain the maximal value of the DVD/LVD ratio from the CTPA axial phase is more intuitive, time-saving and labor-saving than the previous approach of obtaining this parameter from evaluating the 4 chamber cardiac level. The CTPA approach is reliable and has strong repeatability.^[39,40] Routine use of this measurement and analysis using the CTPA axial phase in clinical practice will add important prognostic information for patients with APE.

Because few studies have investigated the correlation between PAh and APE in the literature, our study had innovatively evaluated PAh and PAV as predictors of APE prognosis. In our study, PAh showed no specificity in the prognosis, severity and classification for predicting the severity and prognosis of APE. It has been pointed out that the more the MPA is stretched, the worse its compliance is.^[41] Moreover, the expansibility of MPA is limited. Pulmonary artery diameter is only a weak predictor of APE severity and prognosis as reported in the literature.^[7,9,42,43] The reason may be that the pulmonary artery is an elastic reservoir vessel with abundant elastic fibers and collagen fibers, but fewer smooth muscle components. The main function of the elastic fiber and vascular smooth muscle is to regulate the diameter of the vascular lumen and to change the volume of blood vessels and local blood pressure by contraction or relaxation, but the ability to regulate the length (or height) of the vascular wall is poor. Therefore, PAh is not an ideal index to predict the prognosis and severity of pulmonary embolism. Although PAh showed no significant role in predicting the prognosis and severity of APE, a significant (P < .05) difference was found in the MPA/AO ratio, MPA, RPA, LPA, and PAV between the control and the APE groups, suggesting that pulmonary embolism can cause significant changes in PAV and relevant diameters, which can be used to diagnose the pulmonary hypertension.^[6,19] Compared with the diameter of the pulmonary artery, the 3-dimensional geometry of the pulmonary artery is more consistent with the PAh to make up for the lack of irregular or asymmetric lumen morphology, and has higher sensitivity and specificity.^[44,45] Nonetheless, further investigation showed that PAV was not significantly different between patients with poor prognosis and good prognosis, indicating that the MPA volume cannot enlarge without limitations and that the MPA compliance is limited. On the other hand, age may also play a role in relation to PAV between patients with poor and good prognosis because the interference of the cardiopulmonary risk factors or diseases in the elder patients cannot be completely excluded. Thus, PAV is a promising tool for the diagnosis of APE, but the predictive value of PAV needs to be further explored.

Clinical scoring is a method to assess APE according to the clinical manifestations and risk factors of the patients. The Wells score is one of the most commonly used scoring system at present. The diagnostic efficacy of the Wells score for APE is better than the Geneva score and more suitable for the characteristics of Chinese population.^[46,47] In this study, the Wells score showed significant differences between control patients and patients with a good or poor prognosis (Table 1), indicating that the Wells score has an important value in evaluating the prognosis of APE.

The PAOI is a quantitative index to assess the obstruction degree of pulmonary embolism. CT PAOI has an important value in evaluating the severity and prognosis of patients with APE, especially in-hospital mortality.^[48] In our study, PAOI was significantly different between the good and poor prognosis groups, suggesting that PAOI is valuable in evaluating the short-term prognosis in patients with APE.

Ventricular septal displacement is an indirect sign of an increase in the right ventricular pressure. In CT and echocardiography, ventricular septal displacement has been proven to be a positive predictor.^[11,14] Van der Meer et al did not show that the MPA/AO ratio (P = .66) and the ventricular septal displacement (P = .20) were significantly associated with the mortality of patients with APE.^[49] Our study found a similar outcome, with no significant impact from the MPA/AO ratio, ventricular septal displacement and vena cava reflux on the poor prognosis of APE patients. It may be because the ventricular septal curvature has good specificity, but the ventricular septal motion changes with the cardiac cycle.^[50] Non-gated CT may produce false negative measurement results, affecting its sensitivity in predicting right ventricular dysfunction.

The outcomes of the subgroup analysis indicated that the prognosis of patients with the central type of APE was worse than that of patients with the peripheral type and that the location of embolism had an important predictive value for the prognosis of patients. The degree of pulmonary obstruction was more severe in patients with central pulmonary embolism than those with peripheral pulmonary embolism, which may more likely change the shape of the right heart. This is mainly because the pulmonary embolism directly affects the right ventricular outflow tract, which will lead to an increase of the right ventricular pressure at the end of contraction. The right ventricle parameters on CT imaging can not only objectively reflect the hemodynamic changes and right heart function changes in APE patients, but also show the severity of pulmonary embolism, which is the key to pathophysiological changes of pulmonary embolism. From the perspective of pathophysiology, the severity of pulmonary embolism depends not only on the size and distribution of thrombus, but also on the potential cardiopulmonary status of the patients.

Some limitations existed in our study, including the retrospective and 1-center design, Chinese patients enrolled only, and a small cohort of patients. The CTPA used non-gated CT acquisition, and the time of image acquisition in the cardiac cycle was inconsistent, which affected our conclusion to a certain extent. Moreover, the current parameter measurement was simple and intuitive, but the blood clot and the chamber volume measurement combined with the laboratory biochemical results might be more valuable in assessing the right ventricular function. In this study, the age of onset of the patients was generally high, and the cardiopulmonary function may inevitably be affected by some basic diseases and complications, which is the main reason for the low AUC value of the RVD/LVD ratio. All these issues may affect the generalization of our outcome.

In conclusion, RVD/LVD and systolic blood pressure are independent predictors of poor prognosis in patients with APE. When LPA is > 2.1cm in the central type of APE or the RVD/LVD ratio is > 1.11, the risk of poor prognosis increases, which can be used as important indicators for predicting prognosis in patients with APE.

Author contributions

Conceptualization: Xin Tian, Cai-Ying Li.

Data curation: Jie Hu, Xin Tian, Xiao-Wei Liu, Ya-Zhen Liu.

Formal analysis: Jie Hu, Xin Tian, Ya-Zhen Liu, Bu-Lang Gao, Cai-Ying Li.

Investigation: Jie Hu, Xiao-Wei Liu, Bu-Lang Gao.

Funding acquisition: Xiao-Wei Liu.

Methodology: Cai-Ying Li.

Project administration: Ya-Zhen Liu.

Supervision: Xiao-Wei Liu, Cai-Ying Li.

Validation: Jie Hu, Xin Tian, Xiao-Wei Liu, Ya-Zhen Liu, Bu-Lang Gao, Cai-Ying Li.

Visualization: Xin Tian.

Writing – original draft: Jie Hu.

Writing – review & editing: Xin Tian, Bu-Lang Gao.