Research ArticleORIGINAL RESEARCH

Lung Ultrasound Performed by Primary Care Physicians for Clinically Suspected Community-Acquired Pneumonia: A Multicenter Prospective Study

Francisco Javier Rodríguez-Contreras, Antonio Calvo-Cebrián, Juncal Díaz-Lázaro, Miguel Cruz-Arnés, Fernando León-Vázquez, María del Carmen Lobón-Agúndez, Francisco Javier Palau-Cuevas, Paloma Henares-García, Fernando Gavilán-Martínez, Sandra Fernández-Plaza and Carmelo Prieto-Zancudo
The Annals of Family Medicine May 2022, 20 (3) 227-236; DOI: https://doi.org/10.1370/afm.2796

Article Figures & Data

Figures

  • Figure 1.
    Figure 1.

    STARD diagram showing flow of patients in the study.

    STARD = Standards for Reporting of Diagnostic Accuracy; US = ultrasound.

    a Defined by the study’s inclusion and exclusion criteria (Table 1). All had a history and physical examination performed by the primary care physician.

    b Study protocol did not require recording of number of potentially eligible patients who were missed (due to patient declining to participate or insufficient time to explain the study).

    c All chest radiographs were performed the same day at the referral hospital. For this study, the chest radiograph result used for analyses was the one in the report from the hospital’s radiology department (interpretation of the radiograph by the primary care physician was not included).

  • Figure 2.
    Figure 2.

    Representative images of various lung ultrasound patterns considered in this study.

    Panel A: Normal lung ultrasound showing the A-lines pattern, with a well-defined pleural line (long arrow) and parallel A-lines (short arrows). Panels B and C: Pathological B-lines patterns showing multiple and separated B-lines (arrows in panel B) and coalescent B-lines (arrows in panel C). Panels D, E, and F: Images of consolidations, evident as subpleural hypoechogenic areas with an echogenic air bronchogram. Panel F also shows a hypoechogenic fluid bronchogram (arrow).

  • Figure 3.
    Figure 3.

    Concordance of lung US and chest radiography results for individual patients, according to type of primary care physician performing US.

    US = ultrasound.

    Notes: Figure shows result of lung US performed by primary care physicians (index test) compared with result of chest radiograph as interpreted by a radiologist (reference test). Each block represents a lung US scan and is color coded to show the test result. Blocks are arranged vertically in chronological order, with the first scan at the bottom. Panel A: Adult patients (aged >14 years) with lung US performed by family physicians. Almost one-half of all false results were accounted for by physicians 19 and 20, who had among the most US training and experience. Panel B: Pediatric patients (aged 5-14 years) with lung US performed by primary care pediatricians. Note that despite the initial clinical suspicion of pneumonia, patients with true-negative results (40% of pediatric results) would not have received antibiotics according to the algorithm shown in Figure 4.

  • Figure 4.
    Figure 4.

    Proposed clinical algorithm to guide decision making when using lung US in patients with suspected CAP in primary care.

    CAP = community-acquired pneumonia; US = ultrasound.

Tables

  • Table 1.

    Inclusion and Exclusion Criteria

    Inclusion criteriaa
    Aged 5 years or older with clinically suspected CAP and either of the following features:
      Feverb characterized by 1 of following:
        >72 hours of fever and cough without any improvement
        Fever >72 hours with purulent sputum
        Fever with ≥1 of following: pleuritic pain; focal or asymmetric auscultation of crackles or hypoventilation; dyspnea (subjective); signs of breathing distress such as tachypnea, retractions, or nasal flaring (objective); SpO2 <95%; hemoptysis
      Isolated fever without focus lasting >4 days
      Reappearance or worsening of fever after clinical improvement of a respiratory condition
    No fever, but presence of 1 of following:
      Cough and purulent sputum lasting >4 days
      Cough with ≥1 of following: pleuritic pain; focal or asymmetric auscultation of crackles or hypoventilation; dyspnea (subjective); signs of breathing distress such tachypnea, retractions, or nasal flaring (objective); SpO2 <95%
      Cough lasting >4 weeks, even as an isolated symptom
      Dyspnea in patients aged >75 years
    Exclusion criteriac
    Hospital admission within past 30 days (ie, possible nosocomial pneumonia instead of CAP)
    Pneumonia already diagnosed during current illness by an imaging test
    Receipt of antibiotics for current illness
    Previous diagnosis of chronic obstructive pulmonary disease
    Previous diagnosis of asthma, or suspicion of asthma attack or bronchial spasm in current illness
    Children with previous diagnosis of recurrent wheezing related to viral infections in whom current illness suggests same diagnosis
    Lung or pleural cancer
    Previous pleurodesis
    Previous thoracic surgery
    Other chronic lung diseases (eg, pulmonary fibrosis)
    Terminal disease (life expectancy <6 months)
    Hemodynamic instability
    Declined lung ultrasound and/or chest radiograph
    Inability to go to the hospital for chest radiograph the same day
    Declined to sign informed consent

    • CAP = community-acquired pneumonia; SpO2 = peripheral oxygen saturation.

    • ↵a Inclusion criteria were selected to avoid inclusion solely for clinical suspicion of pneumonia, a term that has greater potential subjectivity in its interpretation.

    • ↵b Temperature ≥38°C not explained by extrathoracic symptoms.

    • ↵c Exclusion criteria were primarily selected to avoid clinical conditions that might confound lung ultrasound findings.

  • Table 2.

    Characteristics of Patients (N = 82) and Primary Care Physicians (N = 28)

    CharacteristicValue
    Patients
    Sex, female, No. (%)42 (51.2)
    Age, median (IQR), y47 (22-60)
    Age by age-groupa
        Pediatric, mean (SD), y8.9 (2.3)
        Adult, mean (SD), y51.3 (16.8)
    Temperature ≥38°C, No. (%)58 (70.7)
    Signs and symptoms, No. (%)
        Cough80 (97.6)
        Expectoration58 (70.7)
        Hemoptysis3 (3.7)
        Pleuritic pain22 (26.8)
        Dyspnea (subjective)27 (32.9)
        Signs of respiratory distress9 (11.0)
    Auscultatory findings, No. (%)b
        Normal16 (19.5)
        Crackles49 (59.8)
        Hypoventilation16 (19.5)
        Rhonchi7 (8.5)
        Wheezing6 (7.3)
    SpO2, median (IQR), %97 (95-98)
    Primary care physicians
    Specialty, No. (%)
        Family physician21 (75)
        Pediatrician7 (25)
    Previous accredited US training, median (IQR), h    85 (49-244)
    Previous US experience, median (IQR), y        4.5 (2-6.75)
    Number of patients recruited per physician, median (IQR) [range]          2 (1-3.8) [1-8]

    • IQR = interquartile range; SpO2 = peripheral oxygen saturation; US = ultrasound.

    • ↵a There were 15 patients in the pediatric age-group (aged 5-14 years) and 67 patients in the adult age-group (aged ≥14 years).

    • ↵b More than 1 finding possible.

  • Table 3.

    Results of Lung US and Chest Radiography

    Imaging Test and ResultPatients, No. (%)
    Lung USa
    Negative (A-lines pattern)29 (35.4)
    Positive53 (64.6)
        ≥1 consolidation measuring >1 cmb25 (30.5)
        B-lines pattern without any consolidation measuring >1 cm28 (34.1)
    Chest radiography
    Negative41 (50.0)
        Normal36 (43.9)
        Peribronchial thickening5 (6.1)
    Positive41 (50.0)
        Alveolar consolidation35 (42.7)
        Interstitial pneumonia4 (4.9)
        Both alveolar and interstitial findings2 (2.4)

    • IQR = interquartile range; US = ultrasound.

    • ↵a Median (IQR) time spent on the scan was 10 min (7-10 min).

    • ↵b Mean (SD) consolidation depth diameter was 2.73 cm (0.86 cm).

  • Table 4.

    Diagnostic Performance of Lung US Compared With Chest Radiography, in Full Cohort and in Subgroups

    Chest Radiography Result
    Lung US ResultPositive, No. (%)Negative, No. (%)Total, No. (%)OR (95% CI) [P Value]Diagnostic Performance of Lung US, Value (95% CI)
    Full cohort
        Positive lung US36 (87.8)17 (41.5)53 (64.6)10.2 (3.3-31.2)Sensitivity: 0.88 (0.75-0.95); specificity: 0.59 (0.43-0.72); PPV: 0.68 (0.55-0.79); NPV: 0.83 (0.66-0.92); positive LR: 2.12 (1.45-3.10); negative LR: 0.21 (0.09-0.49)
        Negative lung US5 (12.2)24 (58.5)29 (35.4)[<.001]
        Total41 (100)41 (100)82 (100)
    Patients’ age group
    Pediatric
        Positive lung US8 (100)1 (14.3)9 (60.0)96a (2.7-3,362)Sensitivity: 1 (0.68-1); specificity: 0.86 (0.49-0.97); PPV: 0.89 (0.57-0.98); NPV: 1 (0.61-1); positive LR: 6.99 (1.14-42.97); negative LR: not calculable
        Negative lung US     0 (0)6 (85.7)6 (40.0)[.001]
        Total8 (100)7 (100)15 (100)
    Adult
        Positive lung US28 (84.8)16 (47.1)44 (65.7)6.3 (2.0-20.2)Sensitivity: 0.85 (0.69-0.93); specificity: 0.53 (0.37-0.69); PPV: 0.64 (0.49-0.76); NPV: 0.78 (0.58-0.90); positive LR: 1.80 (1.23-2.65); negative LR: 0.29 (0.12-0.68)
        Negative lung US5 (15.2)18 (52.9)23 (34.3)[.001]
        Total33 (100)34 (100)67 (100)
    Physicians’ accredited US training time
    <100 hours
        Positive lung US18 (90.0)6 (35.3)24 (64.9)16.5 (2.8-96.7)Sensitivity: 0.90 (0.70-0.97); specificity: 0.65 (0.41-0.83); PPV: 0.75 (0.55-0.88); NPV: 0.85 (0.58-0.96); positive LR: 2.55 (1.32-4.93); negative LR: 0.16 (0.04-0.60)
        Negative lung US2 (10.0)11 (64.7)13 (35.1)[.001]
        Total20 (100)17 (100)37 (100)
    ≥100 hours
        Positive lung US18 (85.7)11 (45.8)29 (64.4)7.1 (1.6-30.6)Sensitivity: 0.86 (0.65-0.95); specificity: 0.54 (0.35-0.72); PPV: 0.32 (0.44-0.77); NPV: 0.81 (0.57-0.93); positive LR: 1.87 (1.17-2.99); negative LR: 0.26 (0.09-0.80)
        Negative lung US3 (14.3)13 (54.2)16 (35.6)[.005]
        Total21 (100)24 (100)45 (100)
    Physicians’ experience using US
    <3 years
        Positive lung US7 (100)3 (37.5)10 (66.7)23.3a (1.0-576.1)Sensitivity: 1 (0.65-1); specificity: 0.63 (0.31-0.86); PPV: 0.70 (0.40-0.89); NPV: 1 (0.57-1); positive LR: 2.67 (1.09-6.52); negative LR: not calculable
        Negative lung US     0 (0)5 (62.5)5 (33.3)[.03]
        Total7 (100)8 (100)15 (100)
    3-6 years
        Positive lung US20 (83.3)5 (35.7)25 (65.8)9 (1.9-41.7)Sensitivity: 0.83 (0.64-0.93); specificity: 0.64 (0.39-0.84); PPV: 0.80 (0.61-0.91); NPV: 0.69 (0.42-0.87); positive LR: 2.33 (1.13-4.82); negative LR: 0.26 (0.10-0.69)
        Negative lung US4 (16.7)9 (64.3)13 (34.2)[.005]
        Total24 (100)14 (100)38 (100)
    >6 years
        Positive lung US9 (90)9 (47.4)18 (62.1)10 (1.1-95.2)Sensitivity: 0.90 (0.60-0.98); specificity: 0.53 (0.32-0.73); PPV: 0.50 (0.29-0.71); NPV: 0.91 (0.62-0.98); positive LR: 1.90 (1.13-3.19); negative LR: 0.19 (0.03-1.28)
        Negative lung US1 (10)10 (52.6)11 (37.9)[.04]
        Total10 (100)19 (100)29 (100)

    • LR = likelihood ratio; NPV = negative predictive value; OR = odds ratio; PPV = positive predictive value; US = ultrasound.

    • ↵a In the 2 cases where a cell contained a 0 value, the OR was calculated by imputing 0.5 for that cell.

Back to top

In this issue

Print
Download PDF
Article Alerts
Email Article
Citation Tools
Get Permissions

Jump to section

Subjects

Keywords