Inappropriate empirical antimicrobial treatment in bloodstream infections patients in the era of multidrug resistance

Tratamiento antimicrobiano empírico inadecuado en pacientes con infecciones del torrente sanguíneo

Terapia Empírica Inapropriada no Tratamento de Infecções de Corrente Sanguínea na Era da Multirresistência Antimicrobiana

Esta obra está bajo una Licencia Creative Commons Atribución 4.0 Internacional.

Pseudomonas aeruginosa is a ubiquitous opportunistic Gram-negative bacillus and is one of the main pathogens responsible for the occurrence of infections related to health care, contributing to the increase in morbidity and mortality rates, hospitalization time and patient costs.^1,2

It mainly affects individuals with comorbidities, such as diabetes, cystic fibrosis and neoplasms; hospitalized and in prolonged use of invasive devices and antimicrobial therapy.^3,4 Among the most common infections are ventilator-associated pneumonia and bloodstream infection (BSI).^3,5,6

Severe infections due to P. aeruginosa, such as BSI, results in higher morbidity and mortality, longer hospitalization, and high costs, especially among hospitalized patients in less developed countries, like Brazil. The problem is greater in large hospitals with many beds complex care levels.^5,6

BSI in the resource-limited countries is largely caused by Gram-negative bacilli multidrug-resistant (MDR) strains that have been increasing in the last few decades^3,5,6,7 and, for this reason, empirical antibiotic treatment of patients with this infection has become a major challenge for physicians, considering that inappropriate empirical antibiotic treatment might be more frequent than desirable.^3,6,7,8,9

Treatment of BSI in countries like Brazil is largely empirical and a challenge for patients with infections for multidrug-resistant (MDR) P. aeruginosa³. These patients had limited treatment options and are highly vulnerable to receiving inappropriate empiric therapy, which contributes to increased length of hospitalization and worst clinical outcomes.^5,6,8

In the current study, we aimed to describe the rates of inappropriate antimicrobial therapy in a cohort of patients with bloodstream infection by carbapenem-resistant P. aeruginosa in the era of widespread antimicrobial resistance. Second, this study explored characteristics associated with the clinical outcomes of these patients in a large Brazilian tertiary-care hospital.

Study design, Patients, and setting

A cohort study was employed to identify the predictors of mortality in the last ten years, and the impact of inappropriate therapy on the outcomes of patients with BSI by carbapenem-resistant P. aeruginosa. Clinical and demographic information about each patient was obtained from hospital records, such as age, gender, days in the hospital, admission to the Intensive Care Unit (ICU), surgery, invasive procedures, such as mechanical ventilation, central venous line, urinary catheter, tracheostomy, hemodialysis, catheter enteral or gastric nutrition, and use of surgical drain, underlying conditions such as diabetes mellitus, chronic renal failure, heart failure, and cancer.

Furthermore, the previous use of one or more of the following antibiotics was verified: piperacillin/tazobactam, gentamicin, amikacin, cefepime, ciprofloxacin, colistin, and carbapenems. These antibiotics exhibit antipseudomonal activity and are part of the local antibiotic policy.

This study was carried out at the Clinical Hospital of the Federal University of Uberlândia, Brazil, a tertiary-care hospital. The clinical microbiology laboratory database was reviewed, and 87 patients who had BSI by carbapenem-resistant P. aeruginosa were selected for the study. Only the first episode of infection was analyzed. Microbial identification and antimicrobial susceptibility test were performed on a Vitek 2 system (bioMérieux Vitek Systems, Hazelwood, MO, USA) for the following antimicrobials: piperacillin/tazobactam, gentamicin, amikacin, cefepime, ciprofloxacin, colistin, imipenem, and meropenem.

Antimicrobial therapy was considered inappropriate when the patient received antimicrobials that did not present "in vitro" activity and/or when treatment was started over 48 hours after the infection diagnosis.¹⁰ Multidrug resistance (MDR) was defined as acquired non-susceptibility to at least one agent belonging to three or more antimicrobial categories.¹¹

Statistical analysis

The Student’s T-test was used to compare continuous variables and x² or Fisher’s exact tests to compare categorical variables. To determine independent risk factors for 30-day mortality, a multiple logistic regression model was used to control the effects of confounding variables. A survival curve was constructed using the Kaplan–Meier method. P values of ≤0.05 were considered statistically significant.

Ethical approval

The research was approved by the Federal University of Uberlandia Committee of Ethics in Research with Human Participants (Approval No. 2.527.621, CAAE: 77541517.9.0000.5152).

Among patients with BSI by carbapenem-resistant P. aeruginosa was observed a high ICU admission rate (63.2%). The previous use of antibiotics and central venous catheter use was common, with 87.3% and 88.5%, respectively. Of 87 eligible patients, 53 (61%) received inappropriate empirical antibiotic therapy of BSI onset, and 67.9% of these patients the infection was caused by P. aeruginosa MDR. The 30-day mortality rate was 57.5% for all patients. However, the 30-day mortality rates were 55% in patients with BSI by P. aeruginosa MDR (33/60). Table 1 summarizes the multivariate analysis of predictors for mortality in BSI by carbapenem-resistant P. aeruginosa.

Table 1
Univariate and multivariate analysis of factors associated with the mortality of 87 patients with bloodstream infection caused by Carbapenem Resistant P aeruginosa

Note: ¹Odds ratio; ²Confidence interval; ³P value; ⁴Multidrug Resistance; *P statistically significant (≤ 0,05).

The mean age among patients who died was 60.2 years (range 9 months to 89 years) and the majority of these patients had more comorbidities than the patients who survived (76% versus 37.8%). Also, a major frequency of invasive devices used was observed among patients with death outcomes within 30 days: CVC (92%), vesical catheter (78%), and mechanical ventilation (76%). The major of patients who died used previous antibiotic therapy (92%) and were submitted to inappropriate empirical antibiotic therapy (84%). Furthermore, inappropriate antibiotic empirical therapy was independently associated with death (P=0.0001).

The mean length of hospitalization was 87 days for survivors and 41.6 days for non-survivors. The Kaplan–Meier cumulative survival estimates (Figure 1) for patients with inappropriate versus appropriate therapy showed that the first group had a lower probability of survival (.=0.0001). The 30-day mortality rate of the first group was 79.2%, while the second group was 23.5%.

Figure 1
Survival curve (30 days) using the Kaplan–Meier method for patients who received antimicrobial appropriate therapy compared with those who received inappropriate therapy

The rapid emergence and spread of carbapenem-resistant P. aeruginosa are a worldwide public health problem, especially in less developed countries like Brazil, where there is a lack of efficient prevention and control actions.^5,12

In Brazil, the mortality rate associated with infection by multidrug-resistant P. aeruginosa is close to 50%, with most patients evolving to death within 30 days after infection.¹³ In countries of the world In the Middle East, mortality reaches 60%,¹⁴ while in Asia this rate is 20%.¹⁵

The development of BSI contributes to the severity of the clinical condition during the hospitalization of the patients who have surgery, are immunocompromised, and use invasive devices. Currently, infection by carbapenem-resistant P. aeruginosa increases the risk of inappropriate empirical antibiotic therapy, frequent in hospitals worldwide, but more frequent in lower and middle-income countries.^5,6,16

This study indicates that older patients, who remained hospitalized for a long period, with some comorbidity and that were in use of inappropriate empirical antibiotic therapy, were associated with worse outcomes. Moreover, our study confirms that a high number of patients with BSI by MDR P. aeruginosa receive inappropriate empirical treatment. According to the literature, the rates of escalation of antibiotic resistance and inadequate therapy considerably increase morbidity and mortality, days of hospitalization, and costs related to the treatment of the infection.^3,7,17

We also identify some factors related to mortality (mean age in years, mean days of hospitalization, cancer, and use of probes for enteral or gastric nutrition) by univariate analysis that concords with previous research conducted mainly in patients with IRAS and BSI by P. aeruginosa.^16,17 Also, we identify that inadequate empirical therapy was associated with a 14-fold increase in mortality. This factor is important because it can be modified.¹⁸

According to the literature, BSI by resistant P. aeruginosa is complex and the impact of inappropriate empirical antibiotic therapy in the mortality of these patients has been a matter of great discussion by researchers and health professionals.^7,16,17 That's why protocols must be developed for each hospital, based on data from active epidemiological surveillance, which allows knowing the predominant species in each hospital unit, as well as the most frequently observed resistance mechanisms.

This study observed that P. aeruginosa causing BSI in this hospital were highly resistant to most antibiotics commonly used in our setting. This trend of resistance could be accounted for by the increasingly empirical, indiscriminate, and intense use of antibiotics. Countless papers showed empiric treatments often fail and increase mortality in these patients.^17,18

In our study, a remarkable number of patients with BSI by carbapenem-resistance P. aeruginosa received inappropriate antibiotic therapy, and this percentage increased when considering MDR P. aeruginosa strains. In light of these results, novel ways of identifying patients with a high risk of MDR P. aeruginosa isolation is mandatory, as well as microbiological diagnosis might help to diminish inappropriate empirical antimicrobial therapy.

Our study had limitations. It was a single-centered, retrospective with small sample size. However, this study generated valuable data regarding infection control aspects, and how it impacts the outcomes of hospitalized patients who had BSI by carbapenem-resistant P. aeruginosa and were treated with inappropriate antibiotic therapy. Results should be validated in other centers and on larger populations.

This study concludes that the rate of inappropriate therapy among ICU patients who developed P. aeruginosa infection was high (61%), being higher among those who died (84%). In addition, the independent risk factors for 30-day mortality were: inappropriate antimicrobial therapy, use of gastric tube and presence of comorbidities, especially cancer.

Ultimately, these findings add some insights about the relationship between higher mortality and inappropriate empirical therapy for patients with BSI by P. aeruginosa. There is a need for better diagnostic tests and infection control programs should focus on de-escalation the antibiotic inappropriate therapy, mainly in BSI caused by carbapenem-resistant P. aeruginosa.

We thank the Hospital of Clinics of Federal University of Uberlândia patients who consented to participate in the study and the staff members for their cooperation.