Response to Matters Arising about the etiology of ventilator-associated pneumonia in the European POS-VAP cohort

Dear Editor,

We would like to thank Ignacio Martin-Loeches and Luis Felipe Reyes [1] for their correspondence concerning our study on incidence, aetiology and outcomes of ventilator-associated pneumonia (VAP) in Europe [2].

It is well-known that micro-organisms responsible for VAP vary by geographic region, patient case mix, duration of hospital stay before onset, and risk factors for colonisation/infection with multi-drug resistant pathogens [3, 4]. Thus, it is not surprising that several large cohort studies describing VAP aetiology rank causative pathogens differently. Generally, most VAP are caused by Gram-negative pathogens, including Pseudomonas aeruginosa, Klebsiella pneumoniae, or Enterobacter or Acinetobacter species, but an important proportion is caused by Gram-positive pathogens, among which Staphylococcus aureus is the most dominant species [5]. In our cohort, Gram-negative pathogens are also implicated in the majority (237/359, 66%) of microbiologically evaluable VAP cases. Nevertheless, at the species level, S. aureus (26.2%) was the most common pathogen (3.9% methicillin resistant S. aureus [MRSA], 17.5% methicillin susceptible S. aureus [MSSA], 4.7% without data), followed by Haemophilus influenzae (16.2%), and P. aeruginosa (15.0%) [2]. Despite Martin-Loeches and Reyes [1] highlighting the importance of P. aeruginosa VAP in the ENIRRI cohort, S. aureus VAP is also frequent in this patient population; combining the reported MRSA (21/556, 3.8%) and MSSA VAP (56/556, 10.1%) gives an overall proportion of 13.9% (95%CI 11.0–16.6), only slightly lower than P. aeruginosa VAP (101/556, 18.9%, 95%CI 15.0-21.3) [6]. Similarly, in the TAVem cohort, 24% of VAP cases could be attributed to P. aeruginosa as well as S. aureus (MRSA 2%, MSSA 22%) [7]. While reported S. aureus VAP proportions vary slightly, all cohorts confirm the increasing importance of MSSA over MRSA, aligned with previous findings in VAP [4] and bloodstream infections in Europe [8].

The higher proportion of P. aeruginosa VAP in TAVem and ENIRRI, could be related to country selection; both include countries from South-America, where the epidemiology is expected to be different from Europe. While country-specific VAP aetiology is not available for these cohorts, an ENIRRI sub-analysis showed that Argentina and Germany report a high proportion of P. aeruginosa nosocomial respiratory tract infections [9], neither are included within POS-VAP. In line with our country-level results, the EU VAP/CAP study [10], including healthcare-associated pneumonia and VAP, shows that S. aureus was the dominant strain in Spain, France, and Belgium, while P. aeruginosa was most common in Italy. Acinetobacter species were the most frequently identified VAP pathogens in Croatia, Romania and Serbia [2]. This highlights the importance of reporting VAP epidemiology by country.

Patient case mix is another important consideration when interpreting reported VAP incidence and aetiology. POS-VAP [2], ENIRRI [6], and TAVem [7] included broadly similar populations, with VAP detected predominantly in males (study proportions: 71–72%), around 60 years of age (study median age: 58–62), with a high proportion of comorbidities, like diabetes (17–19%), and severity scores indicating a high predicted mortality risk (40–46%). Reasons for admission, however, differed. While POS-VAP predominantly included mixed ICUs, 29.5% of ENIRRI VAP cases were identified in medical ICUs [6], where patients are more likely to have prior healthcare exposure. In POS-VAP, important admission diagnoses among VAP cases included stroke (12.4%), trauma (11.2%), and traumatic brain injury (9.2%) [2], conditions associated with an increased risk of S. aureus and H. influenzae VAP [4, 11, 12]. Consistent with this (Table 1), our data shows that S. aureus is the highest ranked VAP pathogen in patients admitted with traumatic brain injury (10/24, 41.7%), stroke (15/49, 30.6%) and trauma (14/48, 29.2%), and it was most prevalent in patients with cardiac decompensation (5/7, 71.4%) or acute myocardial infarction (4/8, 50%). In contrast, P. aeruginosa predominated for more rare diagnoses (‘other primary diagnosis’ 11/38, 28.9%), cardiovascular surgery (4/14, 28.6%), and sepsis/septic shock (4/18, 22.2%). In ENIRRI [6], 26.6% of VAP patients had septic shock on admission, compared with 8% in POS-VAP [2]. Together with the higher proportion of medical ICU patients, this may partly explain the higher proportion of P. aeruginosa VAP in ENIRRI.

Table 1.

Top three pathogens associated with ventilator-associated pneumonia, stratified by admission diagnosis (N=number of microbiologically evaluable [MBE] patients with this diagnosis), among patients in the POS-VAP cohort with microbiological documentation (MBE N=359, n=512 unique pathogen-patient combinations)

A difference in pathogen distribution has also been described for early (≤ 4 days) versus late VAP [13]. As previously reported [2], S. aureus was the most common pathogen in both groups (28.5% and 25.0%, respectively), but the proportion of P. aeruginosa increased from 8.1% to 18.6%, comparing early versus late VAP. As POS-VAP only included first VAP episodes during the first period of invasive mechanical ventilation (IMV), a larger proportion of early VAP cases (31.4%) may have been recorded compared to other VAP cohorts, increasing the importance of S. aureus VAP. Unfortunately, the proportion of early and late VAP is rarely reported in other studies, and it is often unclear how repeated episodes per patient are accounted for. Our stringent criteria avoided double counting of multiple VAP episodes or recurrences in the same patient. As such, POS-VAP data should be considered reflective of VAP occurrence in ICU patients during their first IMV episode.

There is no uniform definition for VAP diagnosis, but most definitions rely on a chest image suggestive of pneumonia combined with the presence of one or more clinical signs and symptoms in patients under IMV for at least 48 h [14]. In addition, microbiological testing can be performed, using blood and/or (invasive) respiratory samples [3]. Within POS-VAP, active VAP screening was encouraged based on FDA criteria, while microbiological confirmation reflected routine clinical practice and was available for 60.7% of VAP cases [2] compared to 75% within ENIRRI [6]. This difference can partly be explained by our stringent 2-day time window around VAP diagnosis date, to ensure identified organisms were associated with the recorded infection. Discrepancies between clinically- and cohort-defined VAP may have also reduced the frequency of microbiological testing, although discordance was low at 5.1%. Microbiological confirmation in POS-VAP was based predominantly on respiratory samples (81.6% endotracheal aspirate [ETA], 15.0% bronchoalveolar lavage [BAL], 5.8% blood culture). In contrast, blood cultures were the main source of pathogen identification within ENIRRI (blood 81.3%, BAL 63.3%, ETA/Sputum 44.8%) [6] and TAVem (blood 74.5%, BAL 28.2%, ETA 68.6%) [7]. Greater reliance on blood cultures likely reflects inclusion of patients with bacteraemic or late-onset VAP, more frequently involving P. aeruginosa, while predominance of ETA samples in POS-VAP may reflect earlier, less severe VAP episodes, although increased detection of respiratory colonisation cannot be ruled out.

High-level VAP surveillance data from multinational cohorts, like POS-VAP, TAVem, or ENIRRI, should not be interpreted as direct guidance for empirical prescribing at the individual patient-level. Hospitals should utilise their own data to inform prescribing guidelines, ideally stratifying microbiological data by ward and infection type to ensure treatment strategies reflect the true ecological context [15]. Therefore, POS-VAP provides feedback to participating sites about their local data and encourages data re-use. In settings where the number of ICU patients, or specifically VAP cases, is low, or routine reporting is limited, national or international surveys can provide supportive information, whereby detailed, stratified results can help determine local applicability.

Prospective cohort data are especially useful for benchmarking across settings and over time, highlighting shifts in incidence, aetiology, or antimicrobial resistance. Differences in VAP incidence, despite similar patient case mix, may indicate gaps in infection prevention and control or antibiotic stewardship strategies. In our cohort [2], adherence to VAP prevention bundle elements ranged from 59.1% (daily sedation vacation) to 93.1% (head-of-bed elevation) and varied between ICUs. Further evaluation of individual bundle elements and their association with VAP risk could help optimise prevention strategies.

To conclude, there is large heterogeneity in VAP aetiology across countries, ICUs, and patient case mix, which complicates the direct comparison of findings between cohorts. Nevertheless, they provide useful evidence to inform pathogen prioritisation for research and development. In addition, data from the POS-VAP network can support site selection and facilitate the design of randomised controlled trials evaluating pathogen-specific preventive or therapeutic interventions in VAP. Based on the findings across studies, it is clear that S. aureus and P. aeruginosa are key targets for improved VAP intervention strategies to enhance clinical outcomes for critically ill patients in Europe.

Author contributions

MdK drafted the manuscript and prepared Table 1. All authors contributed equally to the revision of the manuscript and approved the final version.

Funding

MdK, HJ, AH, CHvW, BF are part of the ECRAID-Base consortium; the ECRAID-Base project has received funding from the European Union’s Horizon 2020 Research and Innovation programme, under Grant Agreement number 965313.

Data availability

The data analysed within is from the prospective, observational POS-VAP study and are not publicly available. However, data access for secondary analyses can be requested from the POS-VAP scientific coordinators (BF, MdK, and CHvW ).

Declarations

Ethics approval and consent to participate

Data from all patients in POS-VAP were collected following the European General Data Protection Regulation and after approval of the POS-VAP protocol by the ethical committees from the participating hospitals and/or national authorities, as applicable. All clinical research has been conducted according to the principles of the Declaration of Helsinki, GCP and local legislations. For patients, ethical procedures followed standard practices for observational studies at each site/country; in most cases, signed informed consent was obtained from patients, or their legal representative. For the remainder of patients, a hospital-wide consent for data use, waiver of consent, or non-objection notice was in place.

Consent for publication

Not applicable.

Competing interests

CHvW reports consultancy fees from MSD/Merck and Sanofi-Pasteur and research grants from MSD/Merck, LimmaTech, BioMérieux and DaVolterra. BF reports consulting fees from Astrazeneca and Aurobac. All other authors declare that they have no competing interests.