Colistin Use Pattern and Nephrotoxicity in Critically Ill Patients with Carbapenem-Resistant Gram-Negative Bacterial Infections. Summary of the Doctoral Thesis

Abstract

Introduction. Colistin is an antibacterial agent of the polymyxin group, which was not used for the treatment of systemic infections for many years due to the risk of neuro- and nephrotoxicity. Nowadays, it is already used in several countries, including Latvia, to treat critically ill patients due to the increased prevalence of multidrug-resistant Gram-negative bacterial infections. Unfortunately, there are currently not many available alternatives for the treatment of these infections, and we must pay special attention to the proper use of existing agents, including the proper dosing of this drug, to minimize the potential development of colistin resistance. This aspect was one of the reasons for our study because, until now, there were no published data on Latvian practices. For a long time, different dosage recommendations were available in different sources of information, especially for patients with impaired kidney function. To achieve the desired effect, colistin is used in high doses. According to the literature, the target therapeutic concentration in the blood plasma could overlap with the toxic concentration. It is important to investigate potentially modifiable risk factors for colistin nephrotoxicity, such as drug interactions with other nephrotoxic drugs. This investigation should not only consider the fact of co-administration with other potentially nephrotoxic substances but also the duration of this co-administration, a factor that has generally been overlooked in previously published studies. Considering the aforementioned importance, the aim of this study is to examine the practice of colistin use and the risk factors for nephrotoxicity in a large tertiary hospital in Latvia. Materials and methods. A retrospective cohort study was conducted, involving adult patients admitted to an intensive care unit between 2015 and 2018, who received parenteral colistin therapy for the treatment of multidrug-resistant Gram-negative bacterial infections for at least 72 hours. The study gathered the following information from medical records: patient demographic data (age, gender), hospitalization details (reason, duration, outcome), clinical information (patient diagnoses), blood biochemistry and microbiological data, details of colistin usage (dosing, duration), concurrent administration of antibacterial agents, and potentially nephrotoxic drugs used in conjunction with colistin. Patients were categorized based on their kidney function at the time of and during the initiation of colistin therapy. The study focused on patients without renal replacement therapy (RRT) at the onset of colistin treatment and those experiencing a rapid increase in serum creatinine levels during colistin therapy. The aim was to investigate the risk factors for acute renal failure in these patients compared to those without colistin-associated renal failure. Results. 111 patients or 117 episodes of colistin use or 1697 days of colistin therapy were analysed. More than half of the patients were men (71 out of 111 or 64 %). The average age of the patients was 61.2 years. The most common diagnoses were pneumonia, subarachnoid haemorrhage, and acute coronary syndrome. The median duration of hospitalization was 44 days and the median day of carbapenem-resistant bacteria isolation was 13. Colistin was most often prescribed for the treatment of pneumonia caused by Acinetobacter baumannii. The median duration of colistin therapy was 11 days (min-max: 3–58) with a median cumulative dose of 78 million units (MU) per treatment episode. In 22 (19 %) episodes of colistin treatment, patients had RRT at the time of colistin initiation and in 26 (22 %) patients had decreased renal function. A generally recommended dosage regimen for colistin is a loading dose equal to 9 MU followed by a daily dose of 9 MU divided into 2–3 administrations. The majority of patients received a loading dose of 9 MU (63.2 %), while the remaining patients either received a reduced loading dose (3–6 MU) or immediately started therapy with a maintenance dose. Patients with impaired renal function had the highest probability of not receiving the standard loading dose (17 out of 26 cases or 65 %) compared to the other groups (p = 0.013). The choice of maintenance dose was also related to the patient's renal functional status. Colistin dosing in 62 % (1047/1697) of all days of colistin therapy was in accordance with recommended doses and in 62 % (13/21) cases. Potential overdoses were most often observed in patients with moderate renal functional status (GFR 30–59 ml/min), in which the patient received a standard colistin dose of 9 MU/day instead of a reduced dose. On the other hand, a potentially lower dose of colistin was found in almost all groups of renal functional status. 24 of 87 cases (27.6 %) had acute kidney injury (AKI) during colistin therapy that met RIFLE criteria (serum creatinine increased by at least 1.5 times from baseline). The median number of days (Q1; Q3) from initiation of colistin therapy to an increase in serum creatinine above 50 % was 8 days and the median cumulative dose of colistin was 68 MU. A statistically significant negative correlation was found between serum creatinine increase (AKI severity) and patient age (Spearman's correlation coefficient −0.578, p = 0.004) as well as baseline creatinine (Spearman's correlation coefficient −0.783, p < 0.001). Patients frequently received at least one potentially nephrotoxic agent (69 of 87, or 79 %) concurrently with colistin, which was a loop diuretic, nonsteroidal anti-inflammatory drug (NSAID), or vancomycin. No associations were found between the use of these agents and the risk of colistin nephrotoxicity. Analysing the potential risk factors of colistin AKI, using univariate logistic regression analysis, it was found that the loading dose of 9 MU increases the risk of AKI (OR = 4.31, p = 0.029), but taking into account a very wide 95 % confidence interval (1.16–16.0), this risk could be lower. On the other hand, the simultaneous use of carbapenem was shown as a protective factor (OR 0.37; CI 0.14–0.97; p = 0.044). Analysing this data with multivariate logistic regression analysis, it can be seen that loading dose remains as a risk factor, but the protective effect of carbapenems was not confirmed. Conclusions. Colistin dosing was consistent with recommended maintenance doses on 62 % of colistin treatment days, and 63 % of colistin treatment episodes were initiated with a standard loading dose. The incidence of colistin AKI in the study center is lower than the published median incidence of colistin AKI in European regions and is equal to 27.6 %. No potentially modifiable risk factors mitigating colistin nephrotoxicity were found, as the risk of colistin AKI is not associated with co-administered potentially nephrotoxic medications or the cumulative dose of colistin. Patients with colistin-induced AKI more often received the loading dose of colistin, and the degree of AKI was more severe in elderly patients with the worst baseline renal functional status.