Editorial Commentary: Pediatric Acute Kidney Injury: Is the Addition of Gentamicin Worth the Risk?

(See the Major Article by Ibrahim et al on pages 1119–24.)

Acute kidney injury (AKI) is not uncommon in children and is associated with significant morbidity and mortality [1–3]. Anti-infectives including aminoglycosides, acyclovir, foscarnet, and amphotericin B are commonly associated with nephrotoxic AKI. Aminoglycoside-induced AKI has been associated with poorer outcomes, even in non–critically ill children [4]. Although the addition of an aminoglycoside occurs in clinical practice for the treatment of infections, compiled evidence demonstrates that this practice is not beneficial for indications such as sepsis and can instead be harmful due to nephrotoxic agent–induced kidney injury. Results from a recent Cochrane analysis discourage the addition of an aminoglycoside to β-lactams in the treatment of sepsis due to the lack of recognized benefit and a significant increase in observed nephrotoxicity [5]. In spite of these data, aminoglycosides continue to be prescribed in such clinical situations. Additionally, the recommendation to consider combination antimicrobial therapy including an aminoglycoside remains embedded in several clinical practice guidelines for diseases such as endocarditis and intravascular catheter–related infections, despite the emerging evidence for harm and uncertainty of benefit [6, 7]. Thus, the critical evaluation of current clinical practice regarding aminoglycoside prescribing continues.

In this issue of Clinical Infectious Diseases, Ibrahim et al report the findings of a retrospective, propensity score–weighted cohort study to compare treatment of ampicillin alone to combination therapy of ampicillin with low-dose gentamicin in the treatment of children with uncomplicated Enterococcus faecalis bacteremia, excluding those with evidence of endocarditis or suppurative thrombophlebitis. Three important outcomes were compared between the 2 treatment groups: duration of bacteremia, frequency of bacteremic relapse, and subsequent risk of the development of AKI.

The primary analysis found that patients receiving combined therapy experienced bloodstream clearance 10 hours more rapidly than those on monotherapy; however, there was no difference in relapse infection rates between treatment groups. Recurrence of bacteremia was not infrequent, occurring 2.5 times more often in those without central line removal regardless of treatment group. Children prescribed gentamicin less frequently had evidence of renal insufficiency or a history of chronic kidney disease at baseline compared with those receiving ampicillin alone before propensity score weighting. The combined therapy patients had twice the odds of developing AKI, and the addition of each nephrotoxic agent significantly increased the likelihood of renal injury. Not every potential nephrotoxic agent was considered, although commonly prescribed medications associated with renal toxicity were accounted for in the analysis.

Patients who received gentamicin in combination with ampicillin were more ill, with an increased percentage requiring intensive care, mechanical ventilation, and vasopressor support. Combination therapy patients were also slightly younger, and more often immunocompromised. The authors attempted to account for these clinical differences, which would increase the risk of AKI, by using propensity score weighting with multivariable logistic regression; however, residual differences between the 2 groups cannot be excluded without performing a randomized clinical trial. Regardless, Ibrahim et al conclude that the addition of gentamicin to ampicillin in the setting of uncomplicated E. faecalis bacteremia likely provides little benefit, with a significant increased risk of renal injury.

The results from this study reiterate the importance of recognizing the risk of inducing pediatric AKI at the point of prescribing a nephrotoxic agent. Medications associated with nephrotoxicity account for approximately 16% of all pediatric inpatient causes of AKI, and an increase in the exposure of nephrotoxic agents increases the risk of developing kidney injury [8]. Increasing severity of AKI is associated with progressively higher mortality and longer lengths of stay [9]. Because children commonly receive nephrotoxic agents in the hospital, predicting those at highest risk of developing nephrotoxicity is challenging, as is determining which of those patients with AKI will suffer long-term sequelae.

A unified definition for pediatric AKI, validated for clinically relevant outcomes, could greatly facilitate the research needed to balance the risk for AKI vs the potential benefits of nephrotoxic therapies. The currently accepted definitions for pediatric AKI, including the pediatric Risk, Injury, Failure, Loss of kidney function, and End-stage kidney disease (pRIFLE) criteria used by Ibrahim et al, are based on consensus criteria and rely on changes in serum creatinine and/or urine output. Defining mild AKI (“risk for renal dysfunction” in the pRIFLE criteria) is problematic in children, as low baseline serum creatinine levels enable very modest absolute changes in the laboratory value to qualify as AKI. On the other hand, increased levels of novel AKI biomarkers such as neutrophil gelatinase–associated lipocalin without elevation in serum creatinine are associated with need for renal replacement therapy, length of stay, and mortality in pediatric patients, indicating that serum creatinine is an insensitive marker for AKI [10]. The potential for novel biomarkers to efficiently identify the onset of kidney injury is encouraging, although currently available tools such as creatinine and urine output remain the gold standard, and delayed detection of AKI frequently occurs [11].

Complications sustained from pediatric AKI are underestimated and not fully understood. Acute renal damage can result in long-term sequelae including residual kidney injury, microalbuminuria, and hypertension [12]. These unintended consequences and the cumulative effect of nephrotoxic agents must be considered at the point of aminoglycoside initiation. Ibrahim et al observed a 63% increase in the odds of developing AKI with the addition of each nephrotoxic agent. Similarly, Moffet et al found that the administration of >2 nephrotoxic agents in children placed them at greater risk of developing AKI [8]. As the vast majority of hospitalized children receive at least 1 nephrotoxic agent, the risk of additional nephrotoxic medications must be considered and balanced against the therapeutic effect prior to prescribing. As for the case of uncomplicated Enterococcus bacteremia, the recognized risk of gentamicin-associated nephrotoxicity appears to exceed the benefit.