Association between severity of pulmonary embolism and health-related quality of life

Abstract

Background

Health-related quality of life (QoL) impairment is common after pulmonary embolism (PE). Whether the severity of the initial PE has an impact on QoL is unknown.

Objectives

To evaluate the association between severity of PE and QoL over time.

Methods

We prospectively assessed PE-specific QoL using the Pulmonary Embolism Quality of Life (lower scores indicate better QoL) questionnaire and generic QoL using the Short Form 36 (higher scores indicate better QoL) questionnaire at baseline and 3 and 12 months in older patients with acute PE. We examined whether QoL differed by PE severity based on hemodynamic status, simplified Pulmonary Embolism Severity Index (sPESI), right ventricular function, and high-sensitivity troponin T in mixed-effects models, adjusting for known QoL predictors after PE.

Results

Among 546 patients with PE (median age, 74 years), severe vs nonsevere PE based on the sPESI was associated with a worse PE-specific (adjusted mean Pulmonary Embolism Quality of Life score difference of 6.1 [95% CI, 2.4-9.8] at baseline, 7.6 [95% CI, 4.0-11.3] at 3 months, and 6.7 [95% CI, 2.9-10.4] at 12 months) and physical generic QoL (adjusted mean Short Form 36 Physical Component Summary score difference of −3.8 [95% CI, −5.5 to −2.1] at baseline, −4.8 [95% CI, −6.4 to −3.1] at 3 months, and −4.1 [95% CI, −5.8 to −2.3] at 12 months). Elevated troponin levels were also associated with lower PE-specific QoL at 3 months and lower physical generic QoL at 3 and 12 months. QoL did not differ by hemodynamic status or right ventricular function.

Conclusion

Severe PE based on the sPESI was consistently associated with worse PE-specific and physical generic QoL over time as compared to nonsevere PE.

1. Introduction

About one-third of patients who survive acute pulmonary embolism (PE) continue to suffer from a reduced health-related quality of life (QoL) at 3 months [1], mostly because of residual dyspnea and exercise intolerance, but consecutive depression/anxiety may also play a role [2–5]. Causally, dyspnea and reduced exercise capacity appear to be predominantly due to obesity, cardiopulmonary comorbidity, smoking, and especially muscle deconditioning rather than residual right ventricular (RV) dysfunction or thrombotic pulmonary vascular obstruction [6–8].

Previous prospective and retrospective studies have identified several independent clinical and social predictors of lower QoL after PE, including advanced age, female sex, several comorbid conditions (eg, obesity and cardiopulmonary disease), and unemployment [1,2,5,9–12]. However, whether the severity of the index PE event per se is related to subsequent QoL remains largely unknown. While 2 studies found an association between markers of PE severity at the time of presentation (ie, a simplified pulmonary embolism severity index [sPESI] of ≥1 points, main pulmonary artery diameter) and a worse QoL [2,11], others did not [1,5,9,12].

However, these studies were limited by small sample sizes [2,5,11,12], a cross-sectional rather than prospective design [5,9,12], a long time lag (>3 years) between the index PE and QoL assessment [5,9,12], or the exclusion of physically unfit patients or those with severe comorbid conditions [11]. We therefore examined in a prospective multicenter cohort of older patients whether PE-specific and generic QoL and dyspnea, one of the most important driving factors of QoL, differ over time between patients who present with severe PE and those who present with nonsevere PE.

2. Methods

2.1. Study population

We analyzed data from the Swiss Venous Thromboembolism Cohort, a prospective multicenter cohort study to assess long-term clinical outcomes and QoL in older patients with acute venous thromboembolism (VTE) from all 5 university and 4 non-university hospitals in Switzerland. Consecutive inpatients and outpatients aged ≥65 years with acute, objectively diagnosed VTE were enrolled and followed up between September 2009 and December 2013. Patients with an inability to provide informed consent (ie, severe dementia), conditions incompatible with follow-up (eg, terminal illness), insufficient ability to speak German or French, thrombosis at a site different from the lower limb, catheter-related thrombosis, or previous enrollment in the cohort were excluded. For the current analysis, we considered patients with PE and those with at least one QoL assessment at 3 or 12 months only. A full description of the study methods and patient characteristics has been published previously [13]. The ethics committees at each site approved the study. All patients provided informed consent.

2.2. Baseline patient data collection

Using standardized data collection forms, trained research nurses prospectively collected data on baseline patient demographic characteristics (age, sex, and body mass index), symptoms of PE (dyspnea, chest pain, and cough), PE characteristics (unprovoked vs provoked, localization, and concomitant deep vein thrombosis), vital signs (arterial oxygen saturation and systolic blood pressure), comorbid conditions (immobilization, active cancer, prior VTE, cardiopulmonary disease, peripheral arterial disease, cerebrovascular disease, pulmonary arterial hypertension, chronic renal disease, diabetes mellitus, acute rheumatic disease, and smoking status), physical activity level, imaging (measures of RV function), laboratory findings (high-sensitivity [hs]-cardiac troponin T), antiplatelet/nonsteroidal antiinflammatory drugs, and PE-related treatments (anticoagulants, thrombolysis, thromboembolectomy, and vena cava filter).

2.3. PE severity measures

We used 4 commonly accepted and readily available measures to assess the severity of PE at the time of presentation: hemodynamic status, the sPESI point score, right-to-left ventricular diameter (RV/LV) ratio, and cardiac troponin levels. These 4 measures are an integral part of the 2019 European Society of Cardiology risk classification algorithm and were shown to be associated with short-term overall and PE-related mortality [14–18]. Severe PE was defined as (1) presence of hemodynamic instability (systolic blood pressure of <90 mm Hg, need for intravenous catecholamines, or cardiopulmonary resuscitation), (2) high-risk PE based on the sPESI (≥1 point), (3) an end-diastolic RV/LV ratio of ≥1 in the apical 4 chamber view on transthoracic echocardiography performed within 2 days after PE diagnosis (or if not available, an RV/LV ratio of ≥1 in the 4 chamber view on computed tomography pulmonary angiography) [19,20], or (4) an elevated hs-troponin T level of >14 ng/L. The sPESI is a clinical score developed and validated to assess the risk of early mortality of PE [21]. It includes the following items: age of >80 years, history of cancer, history of chronic cardiopulmonary disease, heart rate of >110/min, systolic blood pressure of <100 mm Hg, and oxygen saturation of <90%. Hs-troponin T was measured by electro-chemiluminescence methods on Cobas e601 analyzers (Elecsys, Roche). The cutoff of 14 ng/L corresponds to the value above the 99th percentile of the healthy population [22].

2.4. Study outcomes

The primary outcome was PE-specific and generic QoL based on the Pulmonary Embolism Quality of Life (PEmb-QoL) questionnaire and the Short Form 36 (SF-36) version 2 questionnaire, respectively, at baseline and 3 and 12 months. The PEmb-QoL was developed and validated to assess disease-specific QoL within the preceding 4 weeks in patients with PE [23,24]. We used validated French and German versions of the PEmb-QoL [25,26]. The questionnaire contains 40 single items rated on a Likert scale, which are pooled into 6 dimensions: frequency of complaints, activities of daily living limitations, work-related problems, social limitations, intensity of complaints, and emotional complaints. Two items are merely descriptive and not scored. The dimension scores were calculated by averaging item scores and then transforming them to a scale from 0 to 100. We then calculated a PEmb-QoL summary score by averaging the dimension scores, with lower score values indicating a better QoL.

Generic health-related QoL was measured using the SF-36 version 2, a well-validated questionnaire that measures QoL and the general perception of health during the preceding 4 weeks [27]. It consists of 36 items grouped into 8 dimensions (physical functioning, social functioning, physical role functioning, emotional role functioning, mental health, vitality, bodily pain, and general health). The scores vary from 0 to 100 for each dimension. The SF-36 provides a Physical Component Summary (PCS) score and a Mental Component Summary (MCS) score, with higher scores indicating better QoL.

The secondary outcome was the intensity of self-reported dyspnea based on a single question from the PEmb-QoL questionnaire (item 8, “How much breathlessness have you experienced in the past four weeks?”). Ratings were classified into 6 categories of increasing dyspnea intensity (none, very slight, slight, quite a bit, serious, and very serious). All QoL questionnaires were self-administered during a face-to-face interview at baseline and 3 and 12 months after the index PE.

2.5. Statistical analysis

Patient baseline characteristics are shown as numbers and percentages for categorical variables and medians with the IQR for continuous variables. We calculated the mean PEmb-QoL summary score and the mean SF-36 PCS and MCS at baseline and 3 and 12 months with their 95% CI. We compared QoL at 3 months vs baseline and at 12 months vs 3 months by determining the mean score differences and 95% CIs for each point in time. We also graphically presented the proportion of patients in each of the 6 dyspnea intensity categories at baseline and 3 and 12 months.

We used mixed-effects models with random intercept for patients to examine whether the previously described 4 measures of PE severity were associated with the PEmb-QoL and the SF-36 PCS and MCS scores. Results were expressed as mean QoL point differences between patients with severe PE and patients with nonsevere PE at baseline and 3 and 12 months, adjusting for factors that were previously found to be associated with QoL after PE (ie, age, sex, body mass index, unprovoked PE, active cancer, bed rest for >72 hours, bone fracture/cast of lower extremities, prior VTE, cardiopulmonary disease, peripheral arterial disease, cerebrovascular disease, diabetes mellitus, and acute rheumatic disease) [1–3,9–11]. We included time as a factorial variable for baseline and 3 and 12 months in the models. As age, active cancer, and cardiopulmonary disease are already part of the sPESI, these variables were omitted from models, including the sPESI.

We calculated odds ratios (ORs) to compare dyspnea intensity based on the 6 dyspnea categories at baseline and 3 and 12 months between patients with severe PE and patients with nonsevere PE based on the 4 severity measures, with the OR indicating the odds of rating a higher dyspnea category. The models were adjusted for the same variables as the QoL analysis and additionally for smoking status. In our primary analysis, missing values of binary baseline variables used to calculate the sPESI were assumed to be normal. PEmb-QoL summary and SF-36 PCS and MCS scores were computed if all mean dimension scores were available, that is, if at least 50% of questions of a given dimension were answered. In a sensitivity analysis, we performed multiple imputations by chained equations to impute missing baseline and outcome data. We generated a total of 50 imputed datasets and analyzed them using Rubin’s rules [28]. All analyses were performed using Stata 17 (Stata Corporation). A P value of <.05 was considered statistically significant.

3. Results

Of 1003 patients enrolled in the Swiss Venous Thromboembolism Cohort, 12 were excluded because they denied the use of their data/ withdrew early from participation, 304 were excluded because they had deep vein thrombosis without PE, and 141 were excluded because they did not complete at least 1 QoL questionnaire, leaving a final study sample of 546 patients with PE (Figure 1). The median age was 74 years (IQR, 69-80 years), 54% were men, 26% had a body mass index of >30 kg/m², and 21% had cardiopulmonary disease (Table 1). Overall, 3% of patients were hemodynamically unstable, 59% had a high-risk sPESI (≥1 points), 20% had an RV/LV ratio of ≥1, and 42% had an hs-troponin T level of >14 ng/L. One patient (0.2%) died within 3 months, and 27 (5%) within 12 months. Five patients (0.9%) were (re-)admitted to the hospital for VTE or bleeding complications within 3 months.

Figure 1. Study flowchart. QoL, quality of life; SWITCO65+, Swiss Venous Thromboembolism Cohort.

Table 1. Baseline patient characteristics.

	Sample (N = 546)	Missing
Characteristic	n (%) or median (IQR)	n (%)
Age >80 y	124 (23)	0(0)
Male sex	294 (54)	0(0)
Body mass index (kg/m2)	27 (25-30)	1 (0)
New or worsening dyspnea	438 (80)	0(0)
Chest pain	267 (49)	0 (0)
New or worsening cough	162 (30)	0(0)
Unprovoked PEa	359 (66)	0 (0)
Central localization of PEb	181 (33)	0(0)
Concomitant deep vein thrombosis	111 (20)	0 (0)
Hypoxemia (arterial oxygen saturation of <90%)	76 (14)	43 (8)
Immobilization during the last 3 mo
Bed rest for >72 h	73 (13)	0 (0)
Bone fracture/cast of lower extremities	10 (2)	0 (0)
Active cancerc	78 (14)	0(0)
Prior venous thromboembolism	159 (29)	0 (0)
Cardiopulmonary diseased	115 (21)	0(0)
Peripheral arterial disease	26 (5)	0 (0)
Cerebrovascular diseasee	43 (8)	0(0)
Known pulmonary arterial hypertensionf	30 (6)	0 (0)
Chronic renal diseaseg	101 (18)	0(0)
Diabetes mellitus	82 (15)	0 (0)
Acute rheumatic diseaseh	10 (2)	0(0)
Current/past smoker	264 (48)	0 (0)
Low physical activity before PE‘	172 (31)	1(0)
Measures of PE severity
Hemodynamic instabilityi	17 (3)	6 (1)
sPESI ≥1 points	324 (59)	0 (0)
RV/LV ratio ≥1k	108 (20)	221 (40)
Hs-troponin T >14 ng/L	228 (42)	68 (12)
Outpatient management	34 (6)	0(0)
Treatments
Anticoagulation only	524 (96)	1 (0)
Systemic or catheter-directed thrombolysis	21 (4)	0 (0)
Surgical thromboembolectomy	1 (0)	0 (0)
Vena cava filter	4(1)	0 (0)

Hs, high-sensitivity; PE, pulmonary embolism; RV/LV, right-to-left ventricular; sPESI, simplified Pulmonary Embolism Severity Index.

^aAbsence of major surgery, estrogen therapy, or immobilization (bed rest for >72 hours, fracture, or cast of the lower extremity, voyage in sitting position for >6 hours) within 3 months of the index PE.

^bClot localized in the pulmonary trunk or main pulmonary arteries.

^cUndergoing chemotherapy, radiotherapy, surgery, and/or palliative care during the last 3 months.

^dChronic or acute heart failure or chronic lung disease.

^ePrior stroke or transient ischemic attack.

^fOf any cause (primary or secondary).

^gKnown history of chronic renal disease (ie, chronic renal failure requiring or not requiring hemodialysis).

^hAcute lumbar pain, sciatica, vertebral compression, or acute arthritis of the lower extremities during the last 3 months.

ⁱMostly lying/sitting activity or avoidance of climbing stairs/carrying light weight (<5 kg).

^jSystolic blood pressure of <90 mm Hg, need for intravenous catecholamines, or cardiopulmonary resuscitation.

^kBased on transthoracic echocardiography within 2 days or, if not available, computed tomography pulmonary angiography.

3.1. QoL and dyspnea over time

Overall, 90%, 90%, and 83% of patients filled out the PEmb-QoL questionnaire at baseline, 3 months, and 12 months, respectively. The proportions of filled-out SF-36 questionnaires were similarly high (92%, 94%, and 83%). The mean PEmb-QoL summary score improved from 37.1 points (95% CI, 35.2-38.9) at baseline to 23.8 (95% CI, 21.9-25.7) at 12 months (Figure 2). Similarly, the mean SF-36 PCS improved from 39.1 points (95% CI, 38.2-40.0) at baseline to 43.6 (95% CI, 42.7-44.5) at 12 months (Figure 2). The most substantial score improvements occurred during the first 3 months, with only modest additional improvements between 3 and 12 months.

Figure 2. Quality of life over time. Mean disease-specific (Pulmonary Embolism Quality of Life questionnaire [PEmb-QoL]) and generic (Short Form 36 questionnaire [SF-36] Physical Component Summary [PCS] and Mental Component Summary [MCS]) quality of life scores with 95% CIs at baseline and 3 and 12 months.

Similarly, data on dyspnea intensity was available in 92%, 93%, and 84% at baseline, 3 months, and 12 months, respectively. Dyspnea intensity decreased particularly during the first 3 months after PE, whereas further improvement was more discreet between 3 and 12 months (Figure 3). For instance, the number of patients who reported serious or very serious dyspnea at baseline decreased from 171 of 501 (34%) to only 35 of 459 (8%) at 12 months.

Figure 3. Intensity of dyspnea over time. Number and proportion of patients in each dyspnea category (none, very slight, slight, quite a bit, serious, and very serious) at baseline and 3 and 12 months.

3.2. QoL and dyspnea by PE severity

Patients with severe PE based on the sPESI had statistically significantly worse (higher) PEmb-QoL summary scores, with an adjusted mean difference between patients with severe PE and patients with nonsevere PE of 6.1 points (95% CI, 2.4-9.8) at baseline, 7.6 points (95% CI, 4.0-11.3) at 3 months, and 6.7 points (95% CI, 2.9-10.4) at 12 months (Table 2). Patients with severe PE based on the sPESI had significantly worse (lower) mean SF-36 PCS scores at all points in time (Table 2). Mean SF-36 MCS scores did not differ by PE severity based on the sPESI over time (Table 2).

Table 2. Quality of life by pulmonary embolism severity at baseline and after 3 and 12 months.

Quality of life measure	PE severity measure, mean score points (95% CI)		Adjusted mean point difference (95% CI)	P-value
	Hemodynamic instability a,b
	Yes	No
Baseline
PEmb-QoL	35.8 (24.8 to 46.8)	37.1 (35.2 to 39.0)	−1.9 (−12.7 to 8.9)	.725
SF-36 PCS	38.0 (32.8 to 43.2)	39.2 (38.3 to 40.1)	−1.8 (−6.7 to 3.1)	.465
SF-36 MCS	47.4 (41.5 to 53.2)	48.2 (47.2 to 49.2)	−0.4 (−6.3 to 5.5)	.889
3 mo
PEmb-QoL	27.5 (17.0 to 38.0)	25.6 (23.8 to 27.5)	1.1 (−9.2 to 11.4)	.840
SF-36 PCS	42.5 (37.5 to 47.4)	43.1 (42.2 to 44.0)	−1.3 (−5.9 to 3.3)	.586
SF-36 MCS	47.9 (42.3 to 53.4)	49.0 (48.0 to 50.0)	−0.9 (−6.4 to 4.7)	.754
12 mo
PEmb-QoL	24.7 (14.0 to 35.4)	23.8 (21.8 to 25.7)	−0.1 (−10.6 to 10.4)	.983
SF-36 PCS	44.0 (38.8 to 49.2)	43.6 (42.6 to 44.5)	0.1 (−4.8 to 5.0)	.974
SF-36 MCS	49.4 (43.5 to 55.2)	49.5 (48.4 to 50.5)	0.3 (−5.6 to 6.2)	.917
	sPESI c
	High-risk	Low-risk
Baseline
PEmb-QoL	39.6 (37.2 to 42.0)	33.4 (30.5 to 36.2)	6.1 (2.4 to 9.8)	.001
SF-36 PCS	37.4 (36.3 to 38.5)	41.6 (40.3 to 43.0)	−3.8 (−5.5 to−2.1)	<.001
SF-36 MCS	47.5 (46.2 to 48.8)	49.0 (47.5 to 50.6)	−1.4 (−3.4 to 0.6)	.173
3 mo
PEmb-QoL	28.9 (26.5 to 31.2)	21.1 (18.3 to 24.0)	7.6 (4.0 to 11.3)	<.001
SF-36 PCS	40.9 (39.8 to 42.1)	46.1 (44.8 to 47.5)	−4.8 (−6.4 to−3.1)	<.001
SF-36 MCS	48.6 (47.3 to 49.8)	49.6 (48.0 to 51.1)	−0.8 (−2.8 to 1.2)	.421
12 mo
PEmb-QoL	26.6 (24.2 to 29.1)	19.7 (16.8 to 22.6)	6.7 (2.9 to 10.4)	<.001
SF-36 PCS	41.7 (40.5 to 42.9)	46.3 (44.9 to 47.7)	−4.1 (−5.8 to−2.3)	<.001
SF-36 MCS	48.6 (47.2 to 49.9)	50.7 (49.1 to 52.3)	−2.1 (−4.2 to 0.0)	.052
	RV/LV ratio b
	≥1	<1
Baseline
PEmb-QoL	36.7 (32.7 to 40.7)	37.0 (34.2 to 39.9)	−0.6 (−5.4 to 4.1)	.792
SF-36 PCS	39.1 (37.2 to 41.1)	39.7 (38.4 to 41.1)	−0.5 (−2.7 to 1.8)	.691
SF-36 MCS	46.5 (44.2 to 48.7)	48.1 (46.5 to 49.7)	−1.6 (−4.3 to 1.1)	.255
3 mo
PEmb-QoL	25.3 (21.3 to 29.3)	25.0 (22.1 to 27.8)	0.2 (−4.6 to 4.9)	.948
SF-36 PCS	44.0 (42.0 to 45.9)	43.4 (42.1 to 44.8)	0.5 (−1.7 to 2.7)	.668
SF-36 MCS	48.1 (45.8 to 50.3)	49.6 (48.0 to 51.2)	−1.7 (−4.4 to 1.0)	.206
12 mo
PEmb-QoL	22.3 (18.2 to 26.4)	23.7 (20.8 to 26.6)	−1.9 (−6.8 to 3.0)	.446
SF-36 PCS	43.8 (41.9 to 45.8)	43.8 (42.3 to 45.2)	0.4 (−1.9 to 2.6)	.756
SF-36 MCS	49.2 (46.9 to 51.5)	49.5 (47.9 to 51.1)	−0.3 (−3.1 to 2.5)	.848
	Hs-troponin T b
	>14 ng/L	≤14 ng/L
Baseline
PEmb-QoL	38.3 (35.4 to 41.2)	36.2 (33.8 to 38.6)	0.2 (−3.5 to 4.0)	.902
SF-36 PCS	37.3 (35.9 to 38.7)	40.4 (39.3 to 41.6)	−1.4 (−3.1 to 0.3)	.117
SF-36 MCS	46.2 (44.6 to 47.7)	49.5 (48.2 to 50.8)	−3.2 (−5.2 to−1.1)	.003
3 mo
PEmb-QoL	29.1 (26.3 to 32.0)	23.3 (20.9 to 25.7)	4.0 (0.3 to 7.7)	.037
SF-36 PCS	40.6 (39.2 to 42.0)	44.8 (43.7 to 45.9)	−2.3 (−4.0 to−0.6)	.008
SF-36 MCS	47.6 (46.0 to 49.1)	50.0 (48.7 to 51.2)	−2.0 (−4.0 to 0.1)	.058
12 mo
PEmb-QoL	26.3 (23.3 to 29.2)	22.0 (19.6 to 24.5)	2.2 (−1.6 to 6.0)	.264
SF-36 PCS	41.3 (39.9 to 42.7)	45.2 (44.0 to 46.4)	−2.2 (−3.9 to−0.4)	.015
SF-36 MCS	48.7 (47.1 to 50.3)	50.1 (48.7 to 51.4)	−1.2 (−3.4 to 0.9)	.259

Hs, high-sensitivity; MCS, Mental Component Summary; PCS, Physical Component Summary; PE, pulmonary embolism; PEmb-QoL, Pulmonary Embolism Quality of Life; RV/LV, right-to-left ventricular; SF-36, Short Form 36; sPESI, simplified Pulmonary Embolism Severity Index.

^aSystolic blood pressure of <90 mm Hg, need for intravenous catecholamines, or cardiopulmonary resuscitation.

^bAdjusted for age, sex, body mass index, unprovoked pulmonary embolism, active cancer, bed rest for >72 hours, bone fracture/cast of lower extremities, prior venous thromboembolism, cardiopulmonary disease, peripheral arterial disease, cerebrovascular disease, diabetes mellitus, and acute rheumatic disease.

^cAdjusted for sex, body mass index, unprovoked pulmonary embolism, bed rest for >72 hours, bone fracture/cast of lower extremities, prior venous thromboembolism, peripheral arterial disease, cerebrovascular disease, diabetes mellitus, and acute rheumatic disease.

Patients with elevated hs-troponin T levels (>14 ng/L) had no differing PEmb-QoL summary scores at baseline and 12 months but significantly worse (higher) scores at 3 months compared to patients with normal troponin levels, with an adjusted mean difference of 4.0 points (95% CI, 0.3-7.7) (Table 2). Similarly, mean PCS scores in patients with an elevated hs-troponin T level were worse (lower) at 3 and 12 months but not at baseline (Table 2). Interestingly, mean MCS scores were also worse (lower) at baseline (adjusted mean difference −3.2 [95% CI, −5.2 to −1.1]) in patients with an elevated troponin level. QoL did not differ by hemodynamic status or RV function (Table 2).

Patients with severe PE based on the sPESI had statistically significantly higher odds of rating a higher dyspnea category at 3 months (OR, 2.1; 95% CI, 1.4-3.3) and 12 months (OR, 2.1; 95% CI, 1.3-3.2) than those of patients with nonsevere PE (Table 3). Severe PE based on the other PE severity measures was not associated with dyspnea intensity at any point in time.

Table 3. Dyspnea by pulmonary embolism severity at baseline and after 3 and 12 months.

Measure of severity	Median (IQR)a		Adjusted ORb (95% CI)	P value
Hemodynamic instabilityc,d	Yes	No
Baseline	4 (2-5)	4 (2-5)	1.0 (0.3-4.2)	.982
3 mo	3 (2-4)	2 (1-4)	1.6 (0.5-5.1)	.466
12 mo	3 (2-4)	2 (1-4)	1.6 (0.5-5.4)	.461
sPESIe	High-risk	Low-risk
Baseline	4 (3-5)	4 (2-5)	1.2 (0.8-1.9)	.404
3 mo	3 (1-4)	2 (1-3)	2.1 (1.4-3.3)	.001
12 mo	3 (1-4)	2 (1-3)	2.1 (1.3-3.2)	.002
RV/LV ratiod	≥1	<1
Baseline	4 (3-5)	4 (2-5)	0.9 (0.5-1.5)	.696
3 mo	3 (2-4)	2 (1-4)	1.3 (0.7-2.1)	.403
12 mo	2 (1-4)	3 (1-4)	1.0 (0.6-1.7)	.969
Hs-troponin Td	>14 ng/L	≤14 ng/L
Baseline	4 (3-5)	4 (2-5)	0.9 (0.6-1.5)	.801
3 mo	3 (1-4)	2 (1-3)	1.4 (0.9-2.1)	.160
12 mo	3 (1-4)	2 (1-3)	1.3 (0.8-2.0)	.328

Hs, high-sensitivity; OR, odds ratio; RV/LV, right-to-left ventricular; sPESI, simplified Pulmonary Embolism Severity Index.

^aDyspnea was rated on a scale from 1 (“none”) to 6 (“very severe”).

^bThe odds of rating a higher dyspnea category.

^cSystolic blood pressure of <90 mm Hg, need for intravenous catecholamines, or cardiopulmonary resuscitation.

^dAdjusted for age, sex, body mass index, unprovoked pulmonary embolism, active cancer, bed rest for >72 hours, bone fracture/cast of lower extremities, prior venous thromboembolism, cardiopulmonary disease, peripheral arterial disease, cerebrovascular disease, diabetes mellitus, acute rheumatic disease, and smoking status.

^eAdjusted for sex, body mass index, unprovoked pulmonary embolism, bed rest for >72 hours, bone fracture/cast of lower extremities, prior venous thromboembolism, peripheral arterial disease, cerebrovascular disease, diabetes mellitus, acute rheumatic disease, and smoking status.

In our primary analysis, missing binary baseline values were assumed to be normal, and missing PEmb-QoL, SF-36 PCS, and MCS scores were disregarded. When we used multiple imputation by chained equations to impute missing values, the results changed only slightly (Supplementary Tables S1 and S2).

4. Discussion

In this prospective cohort of older patients with acute symptomatic PE, physical generic and PE-specific QoL improved over time, in particular during the first 3 months after PE, as shown in previous studies [1,10,11]. The mean baseline SF-36 PCS score in our patients was substantially worse than the norms for a general Swiss population aged >65 years [29], but by month 3, mean scores had improved to the level of healthy population norms. In contrast, the mean SF-36 MCS score did not improve over time and remained worse than the Swiss health population norm [29].

Patients with severe PE based on the sPESI not only had a statistically significantly worse physical generic and PE-specific QoL up to 12 months but were also more likely to have dyspnea at 3 and 12 months than those with nonsevere PE. Based on a minimal clinically important difference (MCID) of 4 points for the PCS score [30], the lower PCS scores at 3 months (−4.8 points) and 12 months (−4.1 points) in patients with severe PE based on the sPESI were clinically relevant. This was not the case for PE-specific QoL based on the PEmb-QoL, where score differences between severe and nonsevere PE at 3 (7.6 points) and 12 months (6.7 points) were below the MCID of 15 points [31]. The latter, however, was established using data from younger, fitter patients without significant comorbid conditions, and it must be left open for discussion whether a 15-point MCID is adequate for older, multimorbid patients. SF-36 MCS scores did not differ by PE severity based on the sPESI.

Albeit to a lesser extent, patients with elevated hs-troponin T levels, a marker of myocardial injury, had a statistically significantly worse physical generic QoL at 3 and 12 months and a worse PE-specific QoL at 3 months. Except for a lower baseline SF-36 MCS score in patients with an elevated troponin level, mental QoL did not differ by troponin. To our knowledge, our study is the first to examine the role of troponin in predicting QoL in PE. How elevated troponin level is related to QoL impairment after PE is unclear. Elevated troponin could be a predictor for prolonged RV dysfunction or residual thrombus burden [32]. However, it is important to note that QoL differences between patients with elevated troponin levels and those with normal troponin levels did not reach the MCIDs.

Overall, an sPESI score of ≥1 points seems to be a better predictor for impaired QoL than elevated troponin level, possibly because the sPESI comprises not only measures of acute cardiopulmonary impairment (ie, tachycardia, arterial hypotension, and hypoxemia), but also patient factors known to be associated with a reduced QoL after PE, that is, age, cardiopulmonary disease, and active cancer [1,2,9–11].

Our results are consistent with findings from a small study demonstrating that patients with severe PE based on the sPESI were more likely to experience limitations in activities of daily living at 6 months than those with nonsevere PE [2]. A prior study found no relationship between 2019 European Society of Cardiology risk categories and QoL based on the PEmb-QoL at 3 and 12 months [1].

We did not find a worse QoL in patients with an RV/LV ratio of ≥1 or in those with hemodynamic instability. Several studies have explored potential baseline imaging predictors of QoL in PE, including clot location/extension, contrast reflux into the inferior vena cava, and the RV/LV ratio, but generally failed to demonstrate a significant association with QoL outcomes [5,9,11,12]. An exception was the main pulmonary artery diameter on computed tomography, which was found to be associated with worse PEmb-QoL scores at 12 months [11]. A higher pulmonary artery systolic pressure on 10-day echocardiography and a reduced percent-predicted peak oxygen uptake (VO₂ peak) at 1 month were also associated with lower QoL [11]. However, both measures are not part of PE routine care and require an additional consultation.

To date, there are no proven interventions to prevent QoL impairment after severe PE. The administration of thrombolysis does not appear to have a relevant impact on QoL or dyspnea after PE [33,34]. As deconditioning plays a major part in exercise impairment after PE [7,8], early exercise-based rehabilitation in patients with PE who are at risk of a persisting QoL decrease could be a promising preventive approach. The only adequately powered randomized trial examining the effect of early, 8-week exercise-based rehabilitation in patients with acute PE did not find an improved QoL at 6 months in the intervention group [35]. However, by excluding patients with severe cardiac disease, chronic obstructive pulmonary disease, and cancer, this trial enrolled a low-risk population (71% had a sPESI of 0 points) who may not benefit from rehabilitation.

Our study has several limitations. First, health-related QoL may differ between older and younger persons, although the interactions might be manifold. While 2 studies indicate a worse QoL with increasing age in patients with PE, Klok et al. [9] found an association between better social functioning and mental health QoL with older age [2,10]. As our sample included patients aged ≥65 years with PE only, our results may not be generalizable to younger patients. Second, given that the sPESI comprises both vital signs at the time of presentation and chronic diseases such as cancer, it most likely reflects not only the direct effect of PE severity on QoL but also effects of chronic conditions. Third, we assessed dyspnea using a single question from the PEmb-QoL instead of a validated questionnaire. Fourth, we did not specifically assess exercise capacity (eg, through a 6-minute walk test), which is another important outcome with direct impact on QoL. Finally, we assumed missing baseline variables to be normal and disregarded missing PEmb-QoL, SF-36 PCS, and SF-36 MCS scores, which may have biased our results in one direction or the other. However, completeness of baseline (except for troponin and RV/RL ratio values where 12% and 40% of data were missing, respectively) and outcomes data was generally high (questionnaire completion of ≥90% at baseline and 3 months and ≥83% at 12 months). When we used multiple imputation for missing data, our results did not change markedly.

In conclusion, physical generic and PE-specific QoL in older patients with acute PE improves particularly during the first 3 months. Patients with severe PE based on an sPESI score of ≥1 points and, to a lesser extent, based on an elevated hs-troponin T level had a worse physical generic and PE-specific QoL at 3 and 12 months than that in those with nonsevere PE. Severe PE based on the sPESI was also associated with a higher likelihood of dyspnea at 3 and 12 months. Whether early exercise-based rehabilitation in patients with severe PE based on the sPESI could improve QoL and reduce symptoms should be examined in a future clinical trial.