Focused cardiopulmonary ultrasound for assessment of dyspnea in a resource-limited setting.

Abstract

Background: The diagnosis and management of acutely dyspneic patients in resource-limited developing world settings poses a particular challenge. Focused cardiopulmonary ultrasound (CPUS) may assist in the emergency diagnosis and management of patients with acute dyspnea by identifying left ventricular systolic dysfunction, pericardial effusion, interstitial pulmonary edema, and pleural effusion. We sought to assess the accuracy of emergency providers performing CPUS after a training intervention in a limited-resource setting; a secondary objective was to assess the ability of CPUS to affect change of clinician diagnostic assessment and acute management in patients presenting with undifferentiated dyspnea.

Methods and results: After a training intervention for Haitian emergency providers, patients with dyspnea presenting urgently to a regional referral center in Haiti underwent a rapid CPUS examination by the treating physician. One hundred seventeen patients (median age of 36 years, 56 % female) were prospectively evaluated with a standardized CPUS exam. Blinded expert review of ultrasound images was performed by two board certified cardiologists and one ultrasound fellowship trained emergency physician. Inter-observer agreement was determined using an agreement coefficient (kappa). Sensitivity and Specificity with confidence intervals were calculated. Pre-test and post-test clinician impressions and management plans were compared to assess for a change. We enrolled 117 patients with undifferentiated dyspnea. Upon expert image review, prevalence of left ventricular systolic dysfunction was 40.2 %, and in those with systolic dysfunction, the average EF was 14 % (±9 %). The parasternal long axis (PLAX) single view was predictive of an overall abnormal echo with PPV of abnormal PLAX 95 % and NPV 93 % of normal PLAX. Weighted kappa for pericardial effusion between the Haitian physicians and two cardiology reviewers was 0.81 (95 % CI 0.75-0.87, p value <0.001) and for ejection fraction was 0.98 (95 % CI 0.98-0.99, p value <0.001). For lung ultrasound, a kappa statistic assessing agreement between the Haitian physician and the EP for pleural effusion was 0.73, and for interstitial syndrome was 0.49. Detailed test characteristics are detailed in Table 3. Overall, there was a change in treating clinician impression in 15.4 % (95 % CI 9-22 %) and change in management in 19.6 % (95 % CI 12-27 %) of patients following CPUS. A significant structural heart disease was common: 48 % of patients were noted to have abnormal right ventricular systolic function, 36 % had at least moderate mitral regurgitation, and 7.7 % had a moderate to large pericardial effusion.

Conclusions: A focused training intervention in CPUS was sufficient for providers in a limited-resource setting to accurately identify left ventricular systolic dysfunction, pericardial effusion, evidence of interstitial syndrome, and pleural effusions in dyspneic patients. Clinicians were able to integrate CPUS into their clinical impressions and management plans and reported a high level of confidence in their ultrasound findings.