Evaluation of Acute Renal Failure After Acinetobacter Baumannii-based Ventilator-associated Event: A Retrospective Cohort Study

Background: Ventilator-associated event (VAE) is the major complication caused mechanical ventilation (MV). We aimed to evaluate whether acute renal failure (ARF) has developed in patients who had been followed-up due to diagnosis of VAE with Acinetobacter baumannii (AcB), and whether renal replacement therapy (RRT) was used, and its relationship with mortality in patients who developed colistin during their treatment. Methods: Retrospective evaluation of the hospital electronic information system records of 2,622 patients were conducted in three years. Patients who had AcB-related VAE and underwent parental colistin treatment were evaluated according to age, gender, diagnosis for intensive care unit (ICU) administration, Acute Physiology and Chronic Health Evaluation (APACHE) II score, colistin dose and treatment duration, requirement for additional antibiotics, total time required for MV, total duration of ICU stay, presence of septic shock, requirement for percutaneous dilatation tracheostomy (PDT), ARF staging according to Kidney Disease Improving Global Outcomes criteria, requirement for RRT and mortality. Results: Total number of VAE cases was 85 (3.19%). AcB-related VAE was detected in 28 patients (32.9%). Bacterial eradication was achieved in 14 patients (50%), clinical response was received in 14 patients (50%), mean colistin dose was 298.2±85.5 mg/day, mean duration of colistin treatment was 14.3±8.6 days. ARF was detected as Stage-I in eight patients (28.6%), Stage-II in four (14.3%) and Stage-III in eight patients (28.6%). There was no difference between patients in need of RRT and those who did not, in terms of age, gender and body mass index. APACHE II score, bacterial eradication, presence of septic shock, clinical response to therapy, daily dose of colistin, duration of colistin treatment, MV duration, PDT requirement and time were similar in groups receiving RRT or not. Conclusion: Colistin treatment of AcB-related VAE caused ARF in 71.5% of the patients and led to serious conditions in 25% of

The increase in the frequency of infections caused by multi-drug resistant gram-negative bacteria, especially Pseudomonas spp. and Acinetobacter spp., and the di culties in their treatment have brought polymyxins back to the agenda (7). Colistin is one of the drugs in this group and the most important side effect of it is nephrotoxicity. The rate of nephrotoxicity caused by colistin in critical patients in the ICU can increase up to 40% (8). This effect is dose dependent and reversible (9). Nephrotoxicity may be severe enough to warrant discontinuation of therapy, or treatment may have to be continued with Renal Replacement Therapy (RRT).
In our study, we aimed to evaluate whether acute renal failure (ARF) developed in patients who were given colistin for the treatment of AcB-related VAE between January 1, 2017 and December 31, 2019 at ICU, and the need for RRT in patients with ARF and mortality rate of the patients.

Methods
Setting and subjects: This study was carried out after the approval of the local ethics committee of the hospital (date: 15.05.2020, number: 2020.05.1.09.041) between 01.01.2017-31.12.2019. The les and the record of the hospital electronic information system Patients who were treated and followed at the ICU of the Anesthesiology and Reanimation Clinic in a tertiary referral hospital were retrospectively evaluated.
The inclusion criteria were being a patient who were diagnosed with AcB-related VAE according to CDC criteria for VAE (10) and developed ARF while being treated with parenteral colistin according to Kidney Disease Improving Global Outcomes (KDIGO) criteria (11) at the ICU ( Figure 1). Patients who were younger than 18 years, who were pregnant, who had previous history ARF, who had chronic renal failure (CRF) and end stage renal disease (ESRD), had intermittent and/or continuous non-invasive mechanical ventilation (NIMV), who had developed pneumonia and/or septic shock and who were treated with colistin via inhalation were excluded from the study.

Microbiological evaluation:
Sputum, endotracheal aspirate (ETA), nonbronchoscopic bronchoalveolar lavage (mini-BAL) or bronchoscopic specimens (bronchoscopic aspirate, BAL) were used for bacteriological culture. For quantitative assessment of lower respiratory tract samples that were obtained via a protected sterile catheter other than sputum, the threshold values were 10 5 cfu / mL for ETA and 10 4 cfu / mL for BAL and mini-BAL.
Colistin administration: By following the normal glomerular ltration rates of the patients and calculating their ideal body weight, colistin (Lixicol®, Pharmaco, Turkey) that included 4.500.000 IU lyophilized injectable powder was administered to the patients at a daily dose of 300 mg twice a day intravenously. One million international units (MIU) of colistin is equivalent to approximately 30 mg dose (12).

Assessment of nephrotoxicity:
After colistin administration on the rst day, renal functions were evaluated daily according to KDIGO criteria (11). Continuous venovenous hemodia ltration (CVVHDF) with regional citrate anticoagulation was applied due to stage 3 ARF, acidosis, electrolyte imbalance, hypervolemia, and hemodynamic instability.

Evaluated parameters:
The parameters recorded and evaluated were ensuring bacterial eradication, presence of septic shock, clinical response (C-Reactive protein (CRP), procalcitonin and leukocyte counts, positive end expiratory pressure (PEEP), FiO 2 , sputum characteristics), colistin dose (mg / day), duration of colistin use (day), the dose and duration of use of additional antibiotics used if necessary, total duration of the mechanical ventilation (MV) (day), percutaneous dilatation tracheostomy (PDT) was performed during the patients' hospitalization in the ICU, if PDT was performed, on which day of hospitalization (day), total length of stay in the ICU (day), need for RRT and eventual discharge or mortality.
The primary endpoint of the study was the RRT requirement according to KDIGO criteria in colistin-dependent ARF used in patients with AcB-related VAE while the secondary endpoint of the study was differences in mortality and discharge rates between patients who underwent RRT and those who did not.

Statistical Analysis:
Mean, standard deviation, median lowest, highest, frequency and ratio values were used in the descriptive statistics of the data. The distribution of variables was measured with the Kolmogorov-Smirnov test. Unpaired sample t-test and Mann-Whitney U test were used to analyze quantitative independent data. Paired sample t-test and Wilcoxon test were used in the analysis of dependent quantitative data. Chi-Square test was used in the analysis of qualitative independent data, while Fisher's Exact test was used when the conditions of Chi-Square test were not met. SPSS 26.0 program was used in the analyzes.

Results
Electronic and written les of a total of 2662 patients were examined. The number of registered VAE cases was 85 (3.19%). Screening results revealed that AcB-related VAE was detected in 28 patients (32.9%; 28/85). The demographic data of the patients are given in Table 1. The most common reasons for hospitalization in the ICU were respiratory failure and diabetes mellitus (DM) and DM-related complications. The types and frequencies of antibiotics combined with colistin in the treatment of AcB-related VAE are shown in Table 2.
While bacterial eradication was achieved in 14 (50%) of the patients, 22 (78.6%) had symptoms of septic shock. While clinical response to colistin treatment was obtained in 14 (50%) patients, the mean colistin dose was 298.2 ± 85.5 mg / day and the mean period of colistin use was 14.3 ± 8.6 days. While the mean duration of MV was 38.7 ± 28.4 days, 15 of the patients (53.6%) required PDT. While the MV period until PDT was 19.7 ± 9.5 days, the duration of ICU hospitalization was found to be 45.9 ± 38.0 days (Table 3). Eight (28.6%) of the patients were classi ed as Stage-I, four (14.3%) as Stage-II, and eight (28.6%) as Stage-III when the ARF was evaluated according to the KDIGO Criteria ( Table 3). Eight of 28 patients (28.6%) had KDIGO Stage-III AKI and needed RRT. There was no difference in age (p = 0.939), gender (p = 0.454) and BMI (p = 0.183) between patients who needed RRT (Group II) and those who did not (Group I) ( Table 4). APACHE II scoring (p = 0.750), bacterial eradication (p = 1.000), presence of septic shock (p = 0.640), clinical response to treatment (p = 1.000), daily colistin dose (p = 0.974) and duration of colistin use between both groups (p = 0.899), the duration of MV (p = 0.899), the need for PDT (p = 0.549), and the duration of MV until PDT (p = 0.609) did not differ (Table 5). There was no difference between the two groups in terms of length of stay in the ICU (p = 0.6290) and mortality (p = 0.053). RRT was more common in post-resuscitation care patients diagnosed with ICU admission (p = 0.038; Table 6).

Discussion
In our nosocomial infection surveillance system, the incidence of VAE was found to be 3.19% (85/2662).
Approximately in 2.4-14.7% cases of pneumonia develop in critically ill patients per thousand ventilator days (13). According to European Union data, the incidence of VAE is around 9% (14). The decrease in the incidence of VAE in recent years may be due to the increased experience of hospital infection control teams and the sensitivity of modernized registry systems. However, great efforts have been made to reduce the incidence of VAE (15,16). In addition, some of these risk factors such as advanced age, respiratory or cardiovascular system disease, organ failure, burns, trauma, acute respiratory distress syndrome (ARDS), gastric colonization, sinusitis, aspiration of high gastric residual volume, which affect the development of VAE, may be present in the admission of patients to the ICU (17,18). Another reason why we have less VAE incidence may be microorganisms that we cannot produce. While microbiological diagnosis results for VAE were around 60-80% in the literature (19,20), bacterial eradication was 50% in our study, and our samples were generally taken from the proximal of the tracheobronchial tree, not the distal (21). Our AcB-related VAE rate was 32.9% of all VAE. We applied combined antibiotic therapy with colistin in all these patients.
The most important side effect of colistin use in the treatment of nosocomial infections associated with AcB is nephrotoxicity (22). The rate of developing nephrotoxicity in patients using colistin was reported to be 6-55% and our results were similar (23). ARF ratio was 71.4% during colistin treatment of the patients in a recent study. Risk factors for developing nephrotoxicity are advanced age, use of other nephrotoxic drugs together with colistin, and the duration and dose of colistin use (24,25). The average age of these patients who t the KDIGO I-II-II staging was 64.5 years and is not advanced. The duration of colistin use was 15 days on average and the average dose was 300 mg / day. We did not discontinue colistin therapy in any of our patients with the effort of completing the treatments. We started RRT in its indication by adjusting the dose according to daily glomerular ltration rate values. Glomeruli are intact in colistin-induced nephrotoxicity, and the side effect is dose-dependent and reversible (26). Therefore, this fact was important for us to maintain this strategy. In studies showing low risk and incidence of nephrotoxicity, ARF de nitions were differentiated and reduced colistin dose regimens were used (27,28). Combination of some drugs with colistin is known to increase the rate of nephrotoxicity such as use of more than two nephrotoxic drugs by Doshi et al. (29) and use of three nephrotoxic drugs by Pogue et al. (30) . Diuretics, angiotensin-converting enzyme inhibitors, contrast agents, aminoglycosides, antimicrobial agents such as amphotericin and rifampicin, calcineurin inhibitors and vasopressors are considered nephrotoxic agents (31). In our study, we could not determine the relation of colistin with nephrotoxicity since we did not make a retrospective analysis of the presence of a second nephrotoxic agent. This drawback is one of the most important limitations of our study. The reason behind the relatively high incidence of nephrotoxicity is the potentiation of nephrotoxicity by other nephrotoxic agents we use.
According to the KDIGO criteria, RRT was required in eight of 20 patients (8/20, 28.6%) in the Stage I-II-II population. All of these 8 patients were mortal. According to this result, colistin-associated nephrotoxicity requiring RRT has a high mortality. All of our patients who needed RRT with CVVHDF were septic shock patients who needed vasopressors. De ning ARF by KDIGO criteria is more predictive than Risk, Injury, Failure, Loss of kidney function, End-stage kidney disease (RIFLE) criteria, which are widely used to predict in-hospital mortality (32). Moving to the de nitions recommended and agreed on by KDIGO criteria is more advantageous compared to RIFLE de nitions with regards to improving patient outcomes with treatment (33). For this reason, in this retrospective study, we used monitoring with KDIGO criteria in order to be both more up-to-date and more guiding. We had some limitations in this study. First, it is the di culty of generalizing the results to different settings and situations resulting from being a single center study. Second, the retrospective design of the study and the surveillance records did not allow us to completely exclude errors between established relationships. The third was the heterogeneity detected during treatment depending on the decisions of the clinical team and the clinical response of the patients. The forth was that the VAE is not suitable de nition for clinical situation and management of the patient (10). The fth was that we do not exclude other agents with nephrotoxicity effects and cannot connect with our results. The sixth was that we could not reach the records on the day of nephrotoxicity and the day of the treatment period when RRT was started. And, nally, since plasma colistin levels could not be measured, we could not explain the relationship between pharmacokinetic, pharmacodynamic and toxicodynamic studies and nephrotoxicity. We warrant that this article is original, has not been formally published in any other peer-reviewed journal, is not under consideration by any other journal and does not infringe any existing copyright or any other third party rights. We are the sole authors of the article and have full authority to enter into this agreement and in granting rights to BMC are not in breach of anyotherobligation. This article contains nothing that is unlawful, libellous, or which would, if published, constitute a breach of contractor of con dence or of commitment given to secrecy. We have taken due care to ensure the integrity of the article. To our -and currently accepted scienti c -knowledge all statements contained in it purporting to be facts are true and any formula or instruction contained in the article will not, if followed accurately, cause any injury, illness or damage to the user.
We agree to the terms of BMC Licence Agreement and Open data policy.
All methods were performed in accordance with the relevant guidelines and regulations by including a statement.

Funding
This study did not receive any funding from any organizations or institutions.

Ethics approval and consent to participate
This study was carried out after the approval of the Ethics Committee of Istanbul Bagcilar Training and Research Hospital (date: 15.05.2020, number: 2020.05.1.09.041).
Informed consent of the all participants were obtained.

Consent for publication
Not applicable.
Availability of data and materials Not applicable. We do not want to share our data sets during review prior the publication because of scienti c plagiarism. Even so, we are ready to present the data sets when you want. The data sets used and/or analysed during the current study available from the corresponding author on reasonable request.

Competing interest
The authors declare they have no competing interests.   Table 6. Mean ICU admission duration, ICU admission diagnosis, mortality differences between groups.