|Home | About | Journals | Submit | Contact Us | Français|
We read with interest the study by Oliveira and colleagues (6). In their article, the authors find a rate of aminoglycoside-associated nephrotoxicity of 58% in intensive care unit (ICU) patients who received gentamicin or amikacin and noted that mortality in these patients was higher than in patients without acute kidney injury (AKI) (45% versus 29%; P = 0.003). The authors conclude that aminoglycosides should be avoided or used with extreme caution in patients with other comorbid conditions that may lead to renal insufficiency. Several important limitations of this study should be addressed.
First, the authors examined all patients receiving aminoglycosides and found that nearly 60% had AKI. However, these patients with AKI had significantly higher rates of multiple risk factors other than aminoglycoside exposure than the non-AKI group. Thus, the AKI group was at increased risk for toxicity due to other factors and not solely aminoglycoside exposure. In addition, the authors state that the aim of their study was to assess the prevalence of aminoglycoside-associated nephrotoxicity, yet there was no effort at determining causality. The fact that on average the AKI occurred on or near day 7 of therapy and was nonoliguric in nature does not confirm a causal association between aminoglycosides and AKI. A more relevant comparator group would have been patients who did not receive aminoglycosides but who were matched to “case” aminoglycoside patients on calendar time, hospital location, and length of stay (4). Such a study design would have allowed comorbidities, severities of illness scores, and effects of exposure to concomitant nephrotoxins to be compared between exposed and nonexposed patients. Furthermore, mortality was increased in the AKI group, which was not surprising, as AKI is associated with increased mortality regardless of etiology. Also of note, serum aminoglycoside troughs were not reported, which is problematic, as the optimization of aminoglycoside pharmacokinetics is known to minimize nephrotoxicity (3, 5).
Additionally, the authors included patients with ≥4 days of aminoglycoside therapy. These inclusion criteria may not accurately reflect aminoglycoside utilization, as patients in the ICU frequently receive empirical aminoglycosides for shorter durations, until culture data are returned.
In the era of increasingly multidrug-resistant Gram-negative bacilli, it is important and often necessary to consider aminoglycosides for treatment. Inappropriate empirical therapy is a known risk factor for mortality in critically ill patients, and in many ICUs Gram-negative bacilli that are β-lactam resistant are often also fluoroquinolone resistant. Multiple studies have demonstrated the ability to improve the appropriateness of empirical β-lactam therapy by ~15% with the addition of an aminoglycoside (1, 2). Therefore, it is crucial that we more accurately determine the risk that aminoglycosides pose to patients.
While we agree that caution should be utilized in prescribing aminoglycosides, we feel that the data presented do not address the question of the independent impact of aminoglycoside exposure on AKI. We also believe that the methods selected did not allow the authors to determine the study's aim, the prevalence of aminoglycoside-induced AKI in ICU patients.
Oliveira and colleagues recently published an interesting retrospective study of ICU patients and aminoglycoside nephrotoxicity (2). In the paper, the authors quote a published aminoglycoside usage rate of 12.1% in ICU patients receiving antibiotics (1). Their study would be considerably strengthened by the addition of a control group consisting of ICU patients receiving antibiotics other than aminoglycosides. Given the above-quoted usage rate, such a matched group should be available in their records. Without a control group, it is not possible to know what percentage of the change in the calculated glomerular filtration rate (cGFR) was due to the aminoglycosides and what percentage was due to other causes. It follows that the calculations regarding factors associated with aminoglycoside toxicity actually address associations with changes in renal function due to all causes in ICU patients. An addendum to the paper would be welcome.
We appreciate the careful reading of our manuscript, the important comments by Peloquin and by Pogue and colleagues, and the opportunity to discuss some important points on aminoglycoside nephrotoxicity.
We agree that the inclusion of a control group would strengthen the results. Indeed, we tried to include one, as suggested in both letters. However, this “retrospective control group” of ICU patients taking other antibiotics was not adequately matched to the group on aminoglycosides, mainly because of the severity of the patients, making this comparison unreliable. So, we recognize that a limitation of our study, as well as of most of the previous ones, is the lack of a control group of patients not exposed to aminoglycosides. It should be acknowledged that the use of other nephrotoxic drugs concurrently with aminoglycosides and that some of the risk factors found, such as hemodynamic instability, are plausible for any population of critically ill patients. Thus, it is possible that the observed prevalence of AKI overestimated the true rate of aminoglycoside-related AKI. On the other hand, the logistic regression analysis was powered and able to identify the independent risk factors associated with AKI development in the population of patients on aminoglycosides, which were diabetes, use of iodinated contrast, hypotension, and the simultaneous use of other nephrotoxic drugs.
Regarding the comments by Pogue et al., we did not state that “aminoglycosides should be avoided or used with extreme caution in patients with other comorbid conditions that may lead to renal insufficiency.” Actually, we specifically suggested that “aminoglycosides should be avoided or used with extreme caution in hypotensive and/or diabetic patients or patients receiving other nephrotoxic drugs, including iodinated contrast, in the ICU.”
We agree that the time interval between the beginning of therapy and AKI diagnosis and the patients’ urinary volume did not confirm a causal association between the antibiotic and the renal injury, and we did not affirm this in the paper. However, the 7-day time frame for AKI development and the fact that almost 100% of the patients presented a nonoliguric AKI are consistent with previous studies and highly suggestive of aminoglycoside-associated nephrotoxicity.
Although other studies have shown increased mortality in patients with AKI, there is little information on the specific influence of aminoglycoside-associated nephrotoxicity on patients’ mortality. So, we believe that it would be useful to call attention to the fact that even a relatively small decrease in the estimated GFR (eGFR) is associated with a poorer prognosis in patients receiving aminoglycoside therapy.
We included patients with ≥4 days of aminoglycoside therapy because shorter time expositions would rarely be associated with AKI development.
Finally, we totally agree that aminoglycosides should be considered for the treatment of severely ill patients. Actually, we even showed that if an appropriate antibiotic dose is used, a decreased baseline GFR is not a risk factor for aminoglycoside-associated nephrotoxicity. On the other hand, the logistic regression clearly pointed for the independent factors associated with the development of AKI in this population of ICU patients receiving aminoglycoside therapy. Taking into consideration that a decreased acute GFR is strongly associated with higher mortality, an increased length of stay, and higher costs for hospitalized patients, other available antibiotic options that are as effective as aminoglycosides, but less nephrotoxic, should be considered for the treatment of patients at high risk for AKI.