Renate

Part 1 of this thesis focused on how conventional diagnostic tests could aid in the diagnosis and prognosis of (suspected) COVID-19, as well as its complications in patients presenting to the emergency department (ED). At the start of the pandemic clinicians turned to computed tomography (CT) to aid rapid diagnosis, as well as risk stratification of suspected COVID-19 patients. The multicenter study in Chapter 2 showed that the CO-RADS (COVID-19 Reporting and Data System) is an excellent tool in differentiating between positive and negative COVID-19 patients at the ED. Furthermore, the degree of pulmonary involvement at ED presentation – quantified by the CT severity score (CTSS) – had a significant positive association with hospital admission, intensive care unit (ICU) admission and 30-day mortality. In Chapter 3, we showed that procalcitonin (PCT) may aid in excluding a bacterial (co)infection in patients with respiratory symptoms presenting to the ED. We found that the PCT cutoff for a bacterial co(infection) can probably be raised to 0.5 μg/L in patients with a proven viral infection. In Chapter 4, we developed and externally validated a prognostication model to predict poor outcome (ICU admission or 30-day mortality) in patients with suspected COVID-19: the COVERED risk score. It showed good discrimination and calibration in both the Dutch derivation cohort and an Italian validation cohort.

In Part 2, we explored the use of point-of-care ultrasound (POCUS) for diagnosis and prognosis in the acute care setting. LUS (Lung Ultrasound) diagnostic properties are better than chest X-ray (CXR) and almost equivalent to chest CT in diagnosing various acute respiratory pathologies including pneumonia and ARDS. The results of Chapter 6 showed that 12-zone LUS had a comparable excellent negative likelihood ratio (NLR) to CT in diagnosing COVID-19 pneumonia. In Chapter 7, we compared different LUS protocols and found the 6-point protocol to be an excellent screening tool due to its high NLR. The findings presented in Chapter 8 show that the semi-quantitative lung ultrasound score (LUSS) at initial ED presentation is associated with poor outcome, admission duration, and disease severity. In Chapter 9, we screened COVID-19 patients admitted to the medical ward for asymptomatic proximal deep venous thrombosis (DVT) with extended compression ultrasound (ECUS), finding a low prevalence (4%), suggesting that systematic screening for asymptomatic DVTs in this setting is not helpful.

Part 3 addressed the use of POCUS in monitoring the COVID-19 disease course in the ICU. Chapter 10 showed excellent agreement between 6-zone and 12-zone LUS scores in both supine and prone positions, indicating no argument for using the more time-consuming 12-zone protocol. In Chapter 12, we found that the LUSS within 24 hours of intubation was associated with successful liberation from mechanical ventilation. Chapter 13 established that serial LUSS is capable of detecting true changes in pulmonary involvement and that a rise in LUSS after 2 weeks is significantly associated with mortality. In Chapter 14, we explored the diagnostic accuracy of multi-organ POCUS for detecting pulmonary embolism (PE) in critically ill patients. We found that a combination of deep vein and cardiac POCUS – rather than any single modality alone – can be of aid in ruling out PE. In Chapter 15, we investigated the evolution of respiratory muscle thickness using POCUS, observing that neuromuscular blocking agent (NMBA) use was associated with muscle loss, while positive end-expiratory pressure (PEEP) was associated with gain in diaphragm thickness.