Tjitske van Engelen

298 Chapter 12 – but data from human studies is lacking [16,17]. We designed a human intervention trial aimed to answer our question whether modulation of the gut microbiome by antibiotics influences lung inflammation in adults with allergic asthma (asthma being the indicator disease for pulmonary inflammatory diseases in general). Our results, presented in Chapter 7, show no effect of short-term antibiotic-mediated gut microbiome disruption on pulmonary inflammation and our data argue against the presence of a gut-lung axis in humans. However, this trial was aimed at shortterm gut microbiome modulation and cannot be compared with data on the role of early-life antibiotics in the onset of pulmonary inflammation or asthma specifically. Also, with such a small and specifically designed human proof-of-concept trial, many questions regarding negative effects of antibiotics on pulmonary inflammation in sepsis management remain unanswered. Diagnosis and clinical care The art of identifying a disease from signs and symptoms and distinguishing it from other possible conditions: diagnosis [18]. To reach a diagnosis, one usually requires information about medical history, current symptoms, physical exam and additional tests such as blood tests and imaging. But above all, one needs a definition of the disease. However seemingly obvious and simple, in practice this can be rather challenging. Pneumonia is defined as a combination of acute signs and symptoms of a lower respiratory tract infection (such as cough, fever, breathlessness and expectoration) and a new pulmonary infiltrate on chest imaging [19]. However, the Emergency Department reality is a patient who coughs every now and then, experiences shortness of breath “but I am a smoker, so what else is new?”, and where the radiologist describes the chest X-ray as “a pulmonary infiltrate on this chest X-ray cannot be excluded”. Does this patient have pneumonia? In medical research, valid and reliable classification of the clinical diagnosis of study participants is a prerequisite[20]. In Chapter 8 we describe our structured approach using a carefully developed reference standard for diagnostic classification. Our structure was designed to include general, high-volume diseases and addresses a common problem: how to classify patients in large scale clinical trials? We designed a classification system that makes use of assessors with increasing levels of medical experience: medical students, residents, and medical specialists. This method proved to be a valid and efficient way to classify the diagnosis of patients suspected of pulmonary disease at the ED. A composite reference, using structured guidelines on disease classification and some form of panel-based consensus, is an accepted and frequently used method in clinical trials[21]. However, specifics are rarely fully described in papers, possibly because of word count limits imposed by journals. This precludes reproducibility and comparison across trials and we hope that our description of methods in a separate paper, serving as a base for all OPTIMACT papers, will inspire other study authors to be equally informative in terms of their description of methods. Chapter 9 found its way into this thesis for the reader to better understand chapters 4, 8, and 10. The methodology, as validated in chapter 8, was widely applied in chapter 9. The OPTIMACT study was designed with the aim of comparing patient outcomes considering the potential replacement of chest X-rays with ultra-low-dose chest CT.

RkJQdWJsaXNoZXIy MTk4NDMw