Accuracy of extended point-of-care lung ultrasound (EPLUS) for aetiological differentiation of lower respiratory tract infections: a prospective cohort study from India.

Abstract

Background: Infectious respiratory diseases significantly cause morbidity and mortality worldwide, particularly in low- and middle-income countries (LMICs) with limited diagnostic resources. This study explored the utility of lung ultrasound (LUS) paired with extra-pulmonary point-of-care ultrasound (POCUS) for differentiating infectious aetiologies in lower respiratory tract infections (LRTI).

Methods: This prospective cohort study was conducted at a tertiary care centre in India. We recruited consenting adults with suspected LRTI who underwent extended point-of-care lung ultrasound (EPLUS). The protocol included thoracic and abdominal views assessing for lung consolidations and B-lines, pleural effusion, pericardial effusion, focal splenic lesions, and abdominal lymphadenopathy and correlated these with patients' final diagnoses.

Results: We included 322 participants in our analysis cohort, which had a notable prevalence of previously existing chronic lung conditions (21%) and an HIV-prevalence of 5%. Infectious lung disease was identified in 48% of patients, comprising 35% with non-TB LRTI and 13% with TB. Non-infectious lung disease was present in 21% of patients. LUS detected consolidations in 75% and >3 B-lines in 72% of participants. LUS findings were mostly non-specific and prevalent across both infectious and non-infectious conditions. Extra-pulmonary ultrasound findings, such as pericardial effusion and splenic lesions, findings common in disseminated tuberculosis, were rare.

Conclusion: The study highlights the high sensitivity of LUS for detecting pulmonary changes but revealed low specificity to differentiate pulmonary conditions, especially in the presence of pre-existing comorbidities. Future research should explore the accuracy of combinations of clinical characteristics and ultrasound findings in algorithmic approaches, which may improve diagnostics in resource-limited settings.