Linking To And Excerpting From Critical Care’s “Diagnostic accuracy of multi-organ point-of-care ultrasound for pulmonary embolism in critically ill patients: a systematic review and meta-analysis”

Abstract

Background

Methods

Results

Discussion

Conclusion

Availability of data and materials

Abstract

Background

The clinical presentation of acute pulmonary embolism (PE) can range from mild symptoms to severe shock, circulatory arrest and even death, thereby presenting with a significant high mortality when undiagnosed. Computed tomography pulmonary angiography (CTPA) is the gold-standard imaging modality for diagnosing PE, however, it has several practical limitations and is not widely available in low-income country settings. In this context, point-of-care ultrasound (POCUS) has emerged as a valuable bedside, non-invasive diagnostic tool. This meta-analysis assesses the accuracy of multi-organ POCUS for diagnosing PE in critical care settings.

Study design and methods

We conducted a systematic search of Pubmed, Embase, Scopus and the Cochrane Library databases for studies comparing multi-organ POCUS with CTPA or ventilation-perfusion scans for PE diagnosis in critical care departments. Two reviewers independently completed search, data abstraction and conducted quality assessment with QUADAS-2 tool. Heterogeneity was examined with I² statistics. We used a bivariate model of random effects to summarize pooled diagnostic odds ratio (DOR), sensitivity, specificity, positive likelihood ratio (PLR), negative likelihood ratio (NLR) and summary receiver operating characteristic (SROC).

Results

Four studies met the inclusion criteria, comprising 594 patients. The mean age of participants ranged from 55.2 to 71 years. Prevalence of PE ranged from 28 to 66.2%. CTPA was the primary reference standard used in most studies. Multi-organ POCUS for PE diagnosis demonstrated a pooled DOR of 25.3 (95% CI 4.43–82.9) with a pooled sensitivity of 0.90 (95% CI 0.85–0.94; I² = 0%) and specificity of 0.69 (95% CI 0.42–0.87; I² = 95%). The PLR was 3.35 (95% CI 1.43–8.02) and the NLR was 0.16 (95% CI 0.08–0.32). The SROC curve showed an AUC of 0.89 (95% CI 0.81–0.94).

Conclusions

Multi-organ POCUS has high diagnostic accuracy for PE diagnosis in critically ill patients. Further research is needed to validated these findings across different patient populations.

PROSPERO registration

CRD42024614328.

Graphical abstract

Background

The clinical presentation of acute pulmonary embolism (PE) can range from mild symptoms to severe shock, cardiac arrest and even death [1,2,3]. Common symptoms include sudden onset dyspnea, chest pain, syncope, and hemoptysis, with dyspnea being the most frequently reported symptom, occurring in a significant majority of patients [2]. This heterogeneity in presentation can be attributed to several factors, such as underlying cause of PE, location and load of thrombus, number of pulmonary lobes affected and the presence of comorbidities [4, 5]. These factos makes the diagnosis of PE challenging, requiring an accurate diagnostic process with a high level of clinical suspicion and a structured stepwise approach. Owing to the nonspecific nature of symptoms and signs, clinical prediction rules such as the Wells score, revised Geneva score, and Pulmonary Embolism Severity Index (PESI) can assist in risk stratification and guide decisions regarding further diagnostic testing [1].

Computed tomographic pulmonary angiography (CTPA) is considered the gold standard for PE diagnosis because of its high sensitivity and specificity [6]. However, the CTPA has several practical limitations in austere scenarios, such as with hemodynamically unstable patients and in limited-resource settings, which can affect its utility. These include high cost, the logistical challenges of transporting an unstable patient to the radiology department, risk of radiation exposure in pregnant patients and limited availability, particularly in low-income countries.Additionally, the use of iodinated contrast material increases the risk of nephrotoxicity and allergic reactions, especially in patients with pre-existing renal impairment or contrast allergies [4].

When CTPA is not feasible, point-of-care ultrasound (POCUS) has demonstrated its usefulness in clinical practice for ruling in or out PE. Each modality—lung ultrasound, leg vein ultrasonography and focused echocardiography—has been independently shown to be useful and accurate method for confirming the diagnosis of PE [7,8,9,10]. However, despite their utility in specific clinical settings, POCUS of each isolated organ system has relatively low sensitivity. As a result, none of these methods alone can reliably rule-out PE. The application of multi-organ POCUS, which combines lung, cardiac and venous ultrasound, has demonstrated increased diagnostic sensitivity for PE in some studies compared with single-organ approaches [11]. A recent meta-analysis [12] evaluated the accuracy of each organ ultrasound in diagnosing PE, but did not assess the performance of multi-organ POCUS approach. Therefore, we conducted a systematic review and meta-analysis on the accuracy of multi-organ POCUS for diagnosing PE in critical care setting.

Methods

This systematic review and meta-analysis was performed and reported in accordance with the Cochrane Collaboration Handbook for Systematic Review of Interventions and the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) Statement guidelines [13, 14]. This meta-analysis involved secondary data from published studies, exempting it from institutional review board approval.

Inclusion criteria

We selected articles assessing the accuracy of combined lung, cardiac and venous (multi-organ) POCUS for the diagnosis of PE. The study population included patients of age ≥ 18 years with suspected PE. Two reference standards for the diagnosis of PE were accepted: CTPA and ventilation/perfusion (V/Q) scan. Moreover, the included studies were required to have a 2 × 2 table of true positive, false negative, true negative, and false positive counts, either extracted from the original article or calculated from other reported information. We excluded preclinical studies, studies including pediatric populations, case reports, conference abstracts, opinion articles, editorials and non-English articles.

Search strategy

A systematic literature search was performed in the following databases: PubMed, MEDLINE/Embase, Scopus and the Cochrane Library. The search strategy included combined terms such as ‘’pulmonary embolism’’, ‘’ultrasound’’ and CTPA/ventilation-perfusion scan-related terms. The detailed string is available in the Supplementary Material. Additionally, a backward search (snowballing) and a forward search (citation-tracking) were conducted for the included articles and relevant literature review. If the required data were not available in the published studies, we contacted the corresponding author to obtain the information.

Study screening and selection

Two authors (R.M. and L.G.) independently screened titles and abstracts and then screened the full texts of the selected articles to identify eligible studies. Any disagreements were resolved through discussion with a third author (R.P). Rayyan.ia [15] software was used to screen, select and exclude duplicate studies.

Data extraction

Each included study was independently scrutinized by two authors (R.M. and L.G.) to obtain the following data: study design, sample size, ultrasonography technique, year, country, median population age, sex proportion, prevalence of PE, diagnosis, reference standard for PE diagnosis, and sensitivity and specificity of multiorgan POCUS for the diagnosis of PE.

Risk of bias assessment

Two authors (R.M. and I.D.) independently performed the Quality Assessment of Diagnostic Accuracy Studies-2 tool (QUADAS-2) [16] to evaluate the risk of bias,which was tailored to suit the review question. Signaling questions were used to assess the following domains: patient selection, index test, reference standard and flow and timing. The risk of bias was assessed across each of the 4 domains and applicability across the first 3 domains. If a study had at least one high-risk domain or two moderate-risk domains, it was rated as having an overall high risk of bias. Disagreements about quality assessment were resolved by consensus by an additional author (L.G.).

Statistical analysis

We performed a meta-analysis of the studies using the reference standard of each study for PE diagnosis. Diagnostic effect measures were obtained from 2 × 2 contingency tables to calculate the sensitivity, specificity, diagnostic odds ratio (DOR), positive likelihood ratio (PLR) and negative likelihood ratio (NLR) with 95% confidence intervals (95% CI). To account for heterogeneity in methodology and demographics across studies, a bivariate random-effects model was used, and forest plots were generated for graphical representation. We constructed summary receiver operating characteristic (SROC) models and calculated the area under the curve (AUC).

We quantified the heterogeneity of the included studies using the inconsistency index (I²). Publication bias was assessed by analyzing funnel plot asymmetry and performing Egger’s test. Statistical significance was assumed for p < 0.05. Statistical analysis was carried out via R software/environment (version 4.4.0, R foundation for Statistical Computing).

Results

Study characteristics

As shown in Fig. 1, the initial search yielded 2688 results. After removing duplicate records and excluding ineligible studies, 16 studies remained for full-text review on the basis of the inclusion criteria. Following further examination, 4 studies were ultimately included, encompassing a total of 594 patients.

Fig 1

PRISMA flow diagram of study screening and selection

The characteristics of the included studies are summarized in Table 1. All studies employed a prospective design. The mean age of participants ranged from 55.2 to 71 years, with the female prevalence varying from 20 to 55%. The prevalence of PE among the study populations ranged from 28 to 66.2%.

CTPA was the primary reference standard used. One study utilized both CTPA and V/Q scan [18]. The main alternative diagnoses observed in addition to PE were, pneumonia, heart failure, COPD, acute coronary syndrome and muscular chest pain. Additionally, one study focused exclusively on patients with COVID-19 [19].

Distinct findings were observed across the included studies for the diagnosis of PE using ultrasound. With respect to lung ultrasound, three studies [11, 18, 19] identified PE by noting subpleural wedge-shaped, triangular, or rounded hypoechoic lesions. One study [19] used two or more subpleural consolidations ≥ 1 cm as the diagnostic criteria for PE. The cardiac ultrasound parameters included in all studies were right ventricle (RV) dilatation, D-shaped interventricular septum and visualization of a thrombus in the right cardiac chambers. Venous ultrasound consistently showed the absence of vein collapse during compression with or without a visible intravascular thrombus, which is indicative of deep vein thrombosis (DVT). A comprehensive list of ultrasound features for PE diagnosis is provided in Table 2.