4.1.1. Diagnosis of CAP
- -
Confirming pneumonia with a chest X-ray on all patients
In our audit, the first point of assessment, which required the confirmation of pneumonia with a chest X-ray for all patients, indicated a high level of guideline adherence; in fact, in almost all patients (99.6%), a chest image (either X-ray or CT scan) was performed at the time of presentation. There was only one patient in which no image was documented in patients’ records. It is worth mentioning that the national CAP guidelines also indicate a possible alternative to the chest X-ray, namely, a thoracic ultrasound in cases where experienced staff are available and X-ray is not possible. In our case, thoracic ultrasound was not performed, most likely due to a lack of availability of adequately trained and experienced staff. The diagnosis in the case not documented by chest imaging was based on the assessment of clinical symptoms, lung auscultation, laboratory results (elevated laboratory values), and the positive results of a urine test (pneumococcal antigen detected). Recent review articles have concluded that urine antigen detection tests have shown a high specificity, suggesting that a positive result indicates the causative pathogen of CAP in clinical practice [
36,
37,
38]. Interestingly, in 13% of cases who underwent a chest image, no infiltrate was recorded. Nevertheless, the diagnosis of pneumonia was made on clinical grounds. In conclusion, while our audit shows a high adherence to chest imaging guidelines for CAP, the specificity of urinary antigen tests and the frequency of cases where chest images did not document any pulmonary infiltrates underscores the need to balance diagnostic rigor with minimizing resource utilization in the management of CAP.
- -
Obtaining two pairs of blood cultures in hospitalized patients
We observed that blood cultures were obtained in a substantial proportion of cases, i.e., in 80% of patients. This finding aligns closely with the results of other audits conducted in similar healthcare settings. For instance, at James Paget University Hospital in the United Kingdom, blood cultures were obtained for 84.2% of cases [
32]. Likewise, at Sligo University Hospital in Ireland, the pre-interventional and post-interventional rates for blood culture collection were 84.4% and 62.5%, respectively [
33]. The most common reason for not obtaining blood cultures, in 20% of cases within our study, despite being recommended in the guidelines, likely includes that antibiotic therapy had already been started. The 10% rate of positive blood cultures in our study is consistent with the published literature and contributes to the ongoing debate about the cost, merit, and clinical implications of obtaining blood cultures in patients hospitalized with CAP [
39,
40,
41].
- -
Urine legionella and pneumococcal antigen testing
The assessment of urinary antigen testing, specifically for the detection of pneumococcal and legionella antigens in hospitalized patients, revealed a high compliance rate, approximately 90%, in the present study compared to other European audits. As evidenced by a study conducted in Italy by Costantini et al. in 2012, adherence to both urinary antigen tests was reported at 55% for all patients [
29]. Conversely, a study in the United Kingdom by Fahimi et al. demonstrated a markedly lower compliance rate, with less than 20% of physicians adhering to the testing protocol [
32]. In Ireland, both pre- and post-intervention results exhibited compliance rates below 20% and up to 40%, respectively [
33]. Among the diagnostic tests for CAP, urine antigen tests have been widely considered useful due to their simplicity of collection and the rapidity of the test results [
36,
42,
43]. However, the variability in terms of guideline compliance highlights the need for harmonizing clinical urinary antigen testing practices.
- -
Obtaining sputum culture if it can be processed within 4 h
It is noticeable that sputum microbiology was only performed on 16.5% of all patients, even though it is recommended to conduct sputum microbiology in all patients admitted to hospital if it can be processed within 4 h. At our hospital, which has an on-site 24-h diagnostic laboratory, a sputum sample can be processed well within 4 h at all times. One possible explanation for the low number of sputum cultures could be the clinical observation that obtaining a sputum sample can be challenging, as it requires patient cooperation and the ability to produce a suitable specimen. Additionally, patients may not fully understand the importance of sputum testing, leading to reluctance or non-cooperation. Other audits also show low numbers of obtaining sputum microbiology. In Ireland, it was just under 20% and approximately 30% in the pre-intervention and post-intervention groups, respectively [
33]. In the audit of El Fahimi et al., the adherence to local guidelines regarding sputum collection was just over 20% in 2015 [
32]. In addition, despite the guidelines suggesting to obtain sputum culture in all patients, the clinical value of sputum cultures in the management of CAP remain controversial [
44]. The published literature suggests that the diagnostic yield of sputum cultures is clearly lower than 50% [
45,
46]. In a recent meta-analysis encompassing 24 studies and involving 4533 adult CAP patients, a bacterial pathogen was identified in only 36% of sputum samples [
47]. However, when good-quality sputum specimens were selected, the test had a summary sensitivity of 0.69 and specificity of 0.91 for detecting Streptococcus pneumoniae.
- -
Performing influenza PCR during influenza season
Throughout the whole 2019 year, influenza PCR was performed on 50% of all patients. The influenza season in the time frame of our study lasted from approximately 1 January 2019 to 20 April 2019 [
48]. During the influenza season, influenza PCR was performed on 72.6% of hospitalized CAP patients. Potential reasons for the lower-than-recommended influenza PCR testing might be related to a low local influenza prevalence or based on clinical presentation, for instance, patients who presented with symptoms that were not strongly indicative of influenza may not have been prioritized for testing. This observation warrants closer examination in the context of optimizing diagnostic practices and the timely detection of influenza cases.
- -
Respiratory multiplex PCR panel only in selected cases
A respiratory panel examination (respiratory multiplex PCR) was performed on 4.3% of patients. Multiplex PCR assay panels allow faster and comprehensive detection of a wide range of clinically relevant markers. In recent years, numerous multiplex PCR assays have been introduced to the market, and the guidelines have recommended the procedure in certain indications [
49]. The hospitalizations who were evaluated in this audit took place in 2019 when this newer form of pathogen identification was not yet recommended in the guidelines.
- -
Peripheral oxygen saturation (SpO2) and respiratory should be measured
The initial assessment of oxygen saturation upon patient admission was consistently recorded for all individuals within our study. However, compliance with the comprehensive documentation of oxygen requirements throughout the hospitalization period was limited. As far as it was documented, 48% of all patients needed oxygen therapy at some point during hospitalization. It is striking that more than half of all patients in this population were tachypneic at the time of presentation. However, the respiratory rate was not documented for 59 patients, so we may have overestimated the prevalence of tachypnoea at presentation, considering that healthcare workers are more likely to assess and document the respiratory rate in tachypneic patients compared to patients with a normal respiratory pattern. The respiratory rate was also moderately documented in other published audits. In clinical audits at the European Gaza Hospital in 2015 and 2016, the respiratory rate was not documented for 73% of patients [
34]. Respiratory rate is relatively easy and quick to assess and it is an important risk parameter for predicting in-hospital mortality [
50]. It is also included in various severity scores such as CRB-65 and qSOFA score. Thus, it should be assessed and documented more comprehensively.
- -
Complete blood count and blood chemistry (including CRP and PCT) should be measured
At our hospital, the inflammation parameters CRP and leucocytes are determined routinely for all patients presenting at the emergency department with respiratory symptoms, so unsurprisingly there were no missing values. On the other hand, procalcitonin is not routinely measured; in our population, PCT was available for only 23% of patients. A meta-analysis by Kamat et al. in 2020 found a pooled sensitivity and specificity of procalcitonin of 0.55 and 0.75, respectively, for detecting bacterial pneumonia [
51]. However, a meta-analysis by Schuetz et al. published in 2018 showed that the measurement of procalcitonin is associated with a reduction in antibiotic exposure as well as a significantly reduced 30-day mortality [
52]. Another meta-analysis by Pepper et al. in 2019 showed an increased survival and shorter antibiotic duration associated with PCT-guided antibiotic discontinuation but noted that there was a low certainty and high risk of bias [
53]. In a 2018 study by Huang et al., PCT-guided therapy did not show a shortened duration of antibiotic therapy compared to usual care [
54]. However, the duration of antibiotic therapy was already very short, with an average of 4.2 days in the PCT-guided group and 4.3 days in the usual group. Our audit showed an average duration of antibiotics of 8.2 days. Thus, implementing a procalcitonin-guided antibiotic treatment at our hospital could potentially facilitate shortening the length of antibiotic therapy and therefore also potential adverse events associated with antibiotics.
- -
Auscultation and percussion should be performed
Pulmonary examination findings were documented for 246 patients (96.9%). In four patients (1.6%), pulmonary examination was not feasible due to an uncooperative patient. In another four patients (1.6%), the examination findings were not documented. Auscultation and percussion are easy assessments that help physicians complete the clinical assessment and differential diagnosis of a patient in addition to laboratory parameters; therefore, it is highly recommended to perform a chest examination during the clinical assessment of any patient with possible CAP.
- -
Arterial blood gas analysis should be performed
Evaluating arterial oxygen levels plays a crucial role in the initial assessment of patients diagnosed with severe CAP. Hypoxemia is associated with potential respiratory failure, ICU admission, and mortality, indicating the severity of organ dysfunction [
55,
56,
57]. Identifying arterial hypoxemia, moreover, has immediate treatment implications, such as supplemental oxygen administration and closer clinical monitoring. Consequently, measuring arterial oxygenation is a crucial quality indicator in the initial management of CAP individuals. In our cohort, a blood gas analysis was performed on 46.5% of all patients and 73% of patients with COPD. These moderate numbers could likely be improved; however, informal discussions with emergency room physicians in our hospital suggest low enthusiasm for blood gas analysis in clinically stable CAP patients who are in no or little respiratory distress.
- -
Additional comments on the diagnosis of CAP
The use of clinical risk scores, such as CRB-65, PSI, SOFA, and qSOFA score, was rarely documented. These scores may help estimate the severity of CAP and distinguish whether patients can be treated as outpatients or should be hospitalized. Moreover, recent studies have demonstrated the significant predictive value of risk scores, not only in terms of mortality but also of other outcomes such as length of hospital stay and rehospitalization [
58,
59,
60,
61,
62,
63]. Improvements in calculating and documenting risk scores are in order. Regarding complications documented in the discharge report, more details are showed in
Table A1 in the appendix.
4.1.2. Therapy of Community-Acquired Pneumonia
- -
Empiric therapy according to the guidelines
In this study, we recorded the administration of 12 distinct antibiotic regimens within our sample. When compared to the Swiss guidelines [
18], in 75% of cases, the antibiotic therapy was administered according to the guidelines. There are interesting discordances in empirical CAP antibiotic therapy guidelines, where the Swiss guidelines advocate ceftriaxone only as an alternative therapy [
18], in contrast to the German guidelines designating it as a first-line antibiotic [
19].
Notably, our study highlighted instances where certain patients received a single dose before the antibiotic therapy was changed. We attribute this most likely to standard practices in the emergency department, where ceftriaxone was initiated as an empirical therapy before a definitive diagnosis was made. Furthermore, our investigation revealed an over-administration of piperacillin/tazobactam. The Swiss guidelines recommend piperacillin/tazobactam only for patients with severe pneumonia (i.e., patients admitted to ICU) and risk factors for resistance, in combination with a macrolide (clarithromycin or azithromycin). In our study, we found that piperacillin/tazobactam was also administered to patients with moderate severity pneumonia (e.g., stable vital signs, no evidence of sepsis, admission to regular hospital ward) without macrolide combination therapy.
Other published audits have found low adherence to local CAP guidelines. In the UK, the National Audit Report from 2018 to 2019 showed an adherence of 58% to local antibiotic guidelines [
64]. An audit of James Paget University Hospital (UK) showed a compliance of 86% with antibiotic-prescribing guidelines [
32]. The audit at Sligo University Hospital (Ireland) showed an increase in overall compliance with local CAP guidelines, from 21.6% to 62.5% (
p < 0.001), after implementing an intervention bundle in 2019 [
33]. At the European Gaza Hospital, 81% of patients received antibiotics that were in line with the local guidelines [
34]. In summary, adherence to the guidelines concerning empiric antibiotic therapy for CAP could be improved at our hospital. Firstly, the use of ceftriaxone should be reduced. Furthermore, the prescription of broad-spectrum antibiotics such as piperacillin/tazobactam should be reserved for cases with documentation of a clear indication. Finally, it is worth considering that patients with severe influenza pneumonia should not receive steroids. To provide more clarity on the impact of diagnostic results on therapy and its duration, we would like to highlight that of the 21 patients in our sample who tested positive for influenza A, only 3 received steroids, due to Addison’s disease, COPD, and asthma, respectively.
- -
Duration of antibiotic therapy: 5 days (at least 2 days after reaching clinical stability), for severe pneumonia, 5–7 days (at least 2 days after reaching clinical stability)
Regarding the duration of antibiotic therapy, the average time of total antibiotic therapy in our dataset was 8.2 days. The Swiss guidelines recommend a duration of 5 days (or 2 days after reaching clinical stability). In severe pneumonia, a duration of 5–7 days is recommended. The average duration of antibiotic therapy in our group was therefore longer than recommended. The published audit from Ireland revealed longer-than-recommended antibiotic duration in their pre-intervention group but shorter-than-recommended duration in the post-intervention group [
33].
- -
Systemic corticosteroids for patients with COPD or asthma with evidence of bronchial obstruction, for 5–7 days
This recommendation is difficult to evaluate retrospectively. Just under 30% of all patients received systemic corticosteroids during their hospital stay. There were several reasons documented why patients received systemic corticosteroids. The most frequent reason was COPD; however, evidence of bronchial obstruction was not documented in any instance. Overall, 75% of patients with COPD received systemic corticosteroids therapy, while 21 patients (28.3%) who received systemic corticosteroids had neither COPD nor asthma. Some patients had reasons unrelated to pneumonia documented for receiving systemic corticosteroids, such as stress prophylaxis during long-term or chronic corticosteroid therapy preceding hospital admission. In our population, severe pneumonia was documented as an indication for steroids in 14 cases. The average length of systemic corticosteroids was just under 5 days, which was shorter compared to the guidelines. Based on our informal discussions with staff internal medicine physicians, we noted considerable awareness to keep steroid treatment in CAP as short as possible. In summary, due to insufficient documentation, we are unable to accurately evaluate the appropriateness of corticosteroid use in the majority of CAP patients in our study.