Discussion

Our study on VAP during two different periods highlighted the following—first, VAP and MDR VAP frequency increased in the COVID-19 period. Second, the onset of VAP after hospitalization was faster during the COVID-19 period. Third, we were unable to identify risk factors associated with MDR VAP, except for female gender, in the COVID-19 period. Fourth, Gram-negative organisms were frequent, with ESBL-producing organisms more common in the COVID-19 period, and finally, the mortality of patients with VAP was extremely high.

Around 22% of ventilated patients in our ICU (80/362) developed VAP during the period of study. That approximately one in five patients develop VAP is not unusual and is similar to other studies. An observational cohort study on VAP in Nepal showed a similar incidence rate of 20% with an incidence density of 16.45 cases per 1,000 ventilator days. ^{Reference Dongol, Kayastha and Maharjan10} In addition, a systematic review and meta-analysis also reported an overall cumulative VAP incidence of 23.5%. ^{Reference Ding, Zhang and Yang11} In our study cohort, patients with COVID-19 infection were more likely to develop VAP compared to patients without COVID-19. The incidence density of VAP during the COVID-19 period was almost double that of the pre-COVID-19 period (19.3 vs 27.8/1,000 ventilator days). This was similarly reported by another study which showed an increased frequency of VAP during the COVID-19 pandemic at 48% compared to 13% prior to COVID-19 with an overall incidence density of 28/1,000 vs 13/1,000 (p = 0.009). ^{Reference Maes, Higginson and Pereira-Dias12} This increased frequency of VAP among COVID-19 patients is perhaps linked to their predilection for bacterial co-infection due to structural lung damage from COVID-19 infection ^{Reference Giacobbe, Battaglini, Enrile and Dentone13} ; moreover, the addition of immunosuppressive agents such as corticosteroids and immunomodulatory medications may also predispose patients to bacterial superinfection in the form of VAP. ^{Reference Boyd, Nseir and Rodriguez14}

We also hypothesize that the global shortage of healthcare supplies during these early stages of the COVID-19 pandemic could have potentially contributed to the higher VAP rate during the COVID-19 period. Although our institution created adaptations to both hospital structure and staff in response to surge capacity, ^{Reference Abad, Lansang and Cordero15} the lack of personal protective equipment such as N-95 masks and gowns, and other critical supplies such as ventilator tubings, which were either recycled or used for longer than recommended, may have increased the potential for cross-contamination. The workforce was also in short supply, and many nurses and staff who were diverted to the ICU had limited training in infection control or were unfamiliar with working in the critical care setting. In addition, compliance with the VAP bundle, which was at 84.6% during the pre-pandemic period, ^{Reference Abad, Formalejo and Mantaring16} was not monitored closely during the pandemic and may have declined further, especially since certain bundle elements (eg, head of bed elevation, readiness to extubate, and oral care) may have been more difficult to assess among critically ill COVID-19 patients who were in isolation and who were either pronated or shielded under an aerosol box.

In our cohort, COVID-19 infection was associated with a faster median onset of VAP compared to non-COVID-19 patients (10 vs 17 days). Increased length of hospital stay is usually reported as a risk factor associated with increased VAP frequency. ^{Reference Wu, Wu, Zhang and Zhong17} However, patients with COVID-19 infection developed VAP despite a shorter duration of hospital stay; this may indicate a sicker cohort of patients (eg, rapid progression of COVID-19 from severe to critical), or perhaps an impaired immune system that specifically predisposed COVID-19 patients to bacterial infection, ^{Reference Russo, Olivadese, Trecarichi and Torti18} explaining the faster onset of VAP among these patients. COVID-19 is known to be associated with cytokine storm and immune dysregulation. Increased cytokine expressions of interleukin-6 (IL-6), interleukin-8 (IL-8), and monocyte chemoattactant protein-1 (MCP-1) are said to be a markers for bacterial and fungal superinfections. ^{Reference Niedźwiedzka-Rystwej, Majchrzak and Kurkowska19} Knowing this, it is not surprising that critically ill patients with COVID-19-associated immune dysregulation are predisposed to superinfection such as VAP. ^{Reference Clancy and Nguyen20,Reference Nag and Kaur21}

In both pre-COVID-19 and COVID-19 studies, ^{Reference Wu, Wu, Zhang and Zhong5,Reference Rouyer, Strazzulla and Youbong22} gender has been identified as an independent risk factor for VAP, with an increase in frequency among males. This is in contrast to our study, where female patients had a higher risk of MDR VAP compared to males. Our finding was similar to another study where there was increased ICU mortality among females (20.5%) compared to males (14.45) (p = 0.113); on logistic regression, female gender was associated with an increased ICU mortality risk [OR 1.775 (CI 1.029–3.062, p = 0.039)]. ^{Reference Koerber, Agaoglu and Bichmann23} However, the relationship of gender with COVID-19 and the subsequent risk of infection such as VAP is controversial. There is heterogeneity of data on the role of gender in COVID-19, and several factors are proposed to influence outcomes, including variations in hormonal and immune system responses, ^{Reference Koerber, Agaoglu and Bichmann23} as well as social behaviors such as smoking. ^{Reference Ma, Zhu, Jiang, Jiang and Xi24} In our cohort, there were no significant differences in baseline demographics of males and females in both pre-COVID-19 and COVID-19 periods, including comorbidities. Thus, no conclusion can be made regarding the role of gender as a risk factor for MDR VAP among COVID-19 patients, but it warrants further exploration.

The most frequently isolated microorganisms during the COVID-19 period in our cohort were Gram-negative organisms (K. pneumoniae and Acinetobacter baumannii). This is similar to other studies, ^{Reference Velásquez-Garcia, Mejia-Sanjuanelo and Viasus25,Reference Karruli, Boccia and Gagliardi28} including in the local setting, where K. pneumoniae was commonly isolated among COVID-19 patients admitted in a tertiary-level government hospital. ^{Reference Salamat, Malundo and Abad26} However, in our cohort, there was an increase in ESBL-producing K. pneumoniae during the COVID-19 period compared to the pre-COVID-19 period. The reasons for this are likely multifactorial, but frequent use of broad-spectrum antimicrobials such as piperacillin-tazobactam and ceftriaxone may have contributed to the emergence of ESBL-producing resistance. ^{Reference Bassetti, Magnasco and Vena27} MRSA frequency was also much higher in the COVID-19 period, which is not surprising since MRSA has been reported to be one of the common causes of bacterial co-infection in patients with COVID-19. ^{Reference Meawed, Ahmed and Mowafy29} In another study, a higher rate of MRSA infection was seen among COVID-19 patients compared to non-COVID-19 patients (65% vs 27.5%). ^{Reference Yohannes, Ahmed and Schelling30} Surprisingly, however, the frequency of carbapenem resistance in our study was lower in the COVID-19 period, contrary to other studies. A possible explanation for this might be due to the local microbial ecology where there is a higher baseline incidence of K. pneumoniae with ESBL-producing resistance. In our institution, microorganisms with carbapenem resistance usually emerge later, after prolonged hospitalization and extensive antimicrobial pressure. Another potential explanation is the small number of patients included in this study which limits inference and generalization. In general, multidrug resistance in COVID-19 patients is likely increased due to overuse of empiric antibiotics, even among those patients with unestablished bacterial pneumonia. ^{Reference Nseir, Martin-Loeches and Povoa31}

VAP has been associated with a significant increase in the 28-day mortality rate among COVID-19 patients. ^{Reference Nseir, Martin-Loeches and Povoa31} This high mortality rate is associated with the increased inflammatory state which predisposes them to immunosuppression. ^{Reference Vacheron, Lepape and Savey32} Disease severity among COVID-19 patients with VAP increases due to an impaired adaptive immune system, especially in patients developing acute respiratory distress syndrome (ARDS). ^{Reference Hue, Beldi-Ferchiou and Bendib33} However, the overall mortality rate in this study (83% and 93%) was high regardless of the study period and higher than reported by other studies, which ranged from 19.4% to 69.2%. ^{Reference Huang, Jiao and Zhang34–Reference Suarez-de-la-Rica, Serrano and De-la-Oliva37} Mortality rate is often influenced by the presence of MDRO infection, septic shock ^{Reference da Silveira, Nedel, Cassol, Pereira, Deutschendorf and Lisboa36} and a higher SOFA score, ^{Reference Feng, Zhou, Zhou, Zou, Wang and Zhang38} all of which were present in this cohort of patients. Other specific risk factors may have contributed to our higher mortality rates, and these need to be identified and studied further.

This study is not without limitations. It is from a single institution and is retrospective, and the small sample size may have underestimated the incidence of MDR VAP and risk factor association. Given the limited number of non-COVID-19 patients admitted during the COVID-19 period, a direct comparison of VAP rates of two different populations within the same time frame was not possible. The results may not be generalizable to another setting. A multicenter study is necessary to further strengthen the factors associated with VAP and to provide more information regarding VAP in other health settings. Nevertheless, it is one of the few studies in the region to characterize and compare VAP among patients in the pre-COVID-19 and COVID-19 periods.

Overall findings in this study highlight that VAP mortality is high. Critically ill COVID-19 patients appear to be at high risk of VAP and MDR VAP compared to non-COVID-19 patients. VAP onset appears to be earlier for COVID-19 patients and should be suspected when patients worsen clinically. Gram-negative pathogens are frequently responsible for VAP and should be considered when deciding empiric therapy. Risk factors associated with MDR VAP should be explored further.