Drug effects on the risk and prognosis of community-acquired pneumonia

This thesis presents a series of studies exploring the hypothesis that both ACE-inhibitors and statins, besides their intended effects, reduce the risk of acquiring pneumonia. Furthermore, as the burden of pneumonia is not only determined by its incidence but also by a high mortality, another widely prescribed class of drugs, antibiotics, is evaluated in relation to the prognosis of pneumonia. In chapter 2 of the thesis the predictive value of antibiotic use prior to hospital admission for pneumonia is examined regarding microbial etiology and clinical outcome. Initial outpatient treatment with beta-lactam antibiotics increased the change of finding atypical pathogens threefold and decreased the probability of pneumococcal pneumonia to a similar degree, whereas, outpatient macrolide treatment was associated with an increased probability of viral pneumonia. Additionally, the possibility that non-responsiveness to outpatient treatment compromises outcome was studied. Prior outpatient therapy showed no association with length of hospital stay nor in-hospital mortality, except for a subgroup of patients with heart failure. These patients had a decreased change of survival after prior outpatient treatment. The use of ACE-inhibitors was examined in patients with diabetes and was associated with a significant dose-dependent protection against pneumonia. The effect was strongest in patients with a recent history of stroke. The latter supports the proposed mechanism of action that ACE-inhibitors prevent aspiration of oropharyngeal flora through improvement of the cough reflex. Additional study showed that the identified ACE insertion/deletion polymorphism is not associated with the risk neither outcome of community-acquired pneumonia. This in contrast with previous findings in Asian populations. reporting the ACE D allele as an independent risk factor for pneumonia. Possibly, the ACE I/D polymorphism is not a functional polymorphism but rather a marker for the true polymorphism for which the linkage differs among ethnic groups. Regarding ACE as a biomarker, serum ACE activity was studied during hospital stay in patients with pneumonia and appeared significantly decreased during the acute phase of pneumonia with return to control level at recovery. Serum ACE activity showed a significant inverse correlation with disease severity as quantified using Acute Physiology Score. The diagnostic and prognostic value of decreased serum ACE activity needs to be further established. Statin use was associated with a significant 50% reduction in pneumonia risk in patients with diabetes. The association was similar for all sorts of statins and independent of trends in prescribing. Regarding this association, the possibility of residual confounding due to selective prescribing of statins to patients with longer life-expectancy could not be excluded. The findings from this thesis were confirmed in four other studies published recently. The last chapter of this thesis focused on the accuracy of International Classification of Diseases codes for identifying patients with pneumonia in administrative databases. The sensitivity for detecting patients with pneumonia in Dutch hospital databases was modest (72%) and dependent on duration of hospital stay. This implies that patients with prolonged and probably complicated hospital-stay are underrepresented in hospital databases. The positive predictive value for coded hospital discharge records as a marker for pneumonia was high (88%)