Use of Aminophylline to Reverse Acute Kidney Injury in Pediatric Critical Care Patients

Abstract

OBJECTIVE

Acute kidney injury (AKI) is a complication encountered in 18% to 51% of pediatric critical care patients admitted for treatment of other primary diagnoses and is an independent risk factor for increased morbidity and mortality. Aminophylline has shown promise as a medication to treat AKI, but published studies have shown conflicting results. Our study seeks to assess the reversal of AKI following the administration of aminophylline in critically ill pediatric patients.

METHODS

We performed a single-institution retrospective chart review of pediatric inpatients who were diagnosed with AKI and subsequently treated with non-continuous dose aminophylline between January 2016 and December 2018. Data were collected beginning 2 days prior to the initial dose of aminophylline through completion of the 5-day aminophylline course.

RESULTS

Nineteen therapies among 17 patients were included in analysis. Twelve of the therapies resulted in resolution of AKI during the study period. We observed urine output increase of 19% (p = 0.0063) on the day following initiation of aminophylline therapy in the subset of patients whose AKI resolved. Trends toward decreased serum creatinine and lower inotropic support were also noted.

CONCLUSIONS

Based on these findings, aminophylline could be considered a potentially effective medication for use as rescue therapy in critically ill children with AKI. Limitations include small study population and retrospective nature. Further research in this area with a larger study population and a randomized control trial would allow for better characterization of the efficacy of aminophylline in reversal of AKI.

Introduction

Acute kidney injury (AKI) is a frequently encountered complication in pediatric critical care settings, with incidence ranging from 18% to 51%.^1–5 Kidneys are sensitive to the decreased perfusion frequently encountered in severe illness, commonly due to sepsis or organ dysfunction,⁶ therefore it is not surprising that critically ill children have a high incidence of AKI. However, AKI has also been noted to be an independent marker for increased morbidity and mortality,^4,5,7–10 suggesting that treatment and prevention of AKI may be an important target for improving overall pediatric intensive care unit (ICU) outcomes; nevertheless, available therapeutic approaches for treatment or reversal of AKI are lacking.

Adenosine receptor antagonists, including theophylline and its prodrug aminophylline, are well-established as treatment for pediatric respiratory conditions such as asthma, due to their bronchodilatory effects; however, their use has been supplanted by new bronchodilators with greater efficacy, less toxicity, and easier administration.¹¹ In addition to its respiratory effects, adenosine has shown clinical benefit in AKI^12–17 in both animal and human studies. Adenosine is a vasoconstrictor of the renal afferent arterioles, an action that leads to reduction in renal blood flow, decreasing glomerular filtration, impairing renal function, and leading to diuretic resistance.^18,19 Aminophylline is a non-selective adenosine receptor antagonist, which has shown promise as a reno-protective agent in the setting of cardiopulmonary bypass,^13,20 neonatal asphyxia,²¹ and calcineurin inhibitor toxicity.²² However, this effect has not been consistently demonstrated. A placebo-controlled randomized controlled trial of 144 children found no effect of aminophylline on prevention of AKI following cardiopulmonary bypass.²³ At our institution, standard therapy for AKI consists of intravenous (IV) loop diuretics (continuous bumetanide and/or intermittent furosemide) with possible addition of intermittent IV chlorothiazide. Aminophylline is routinely used as an adjunctive rescue therapy in the treatment of AKI in our critical care units, particularly in the cardiovascular ICU. The primary purpose of this study was to evaluate the resolution of AKI in pediatric critical care patients when aminophylline was administered. Secondary outcomes included decreased use of inotropes assessed by a vasoactive inotrope score modified to include continuous calcium infusion (VISCa), decreased serum creatinine, and improved urine output (UOP).

Materials and Methods

This was a retrospective case series at a single, free-standing, tertiary-care pediatric hospital affiliated with the University of Tennessee Health Science Center in Memphis, Tennessee. Data were extracted from the Le Bonheur Children's Hospital electronic medical record using ICD-10²⁴ codes for AKI (N17.0, N17.1, N17.2, N17.8, N17.9). All inpatients who were treated with aminophyl-line between January 2016 and December 2018 were assessed.

Aminophylline Therapy. Data were reviewed for consistency of aminophylline administration. A standard course of aminophylline for treatment of AKI was defined as approximately 5 mg/kg/day (max 300 mg/day) intravenously infused over 20 to 30 minutes either daily or divided twice daily for 5 consecutive days.²⁰ Patients who received this standard course of aminophylline (5 mg/kg/day ± 10%) were eligible and were included in this study; patients whose daily dose deviated by more than 10% from the standard dosing were excluded. Patients receiving aminophylline for treatment of status asthmaticus or as a continuous infusion for pulmonary vasodilatory effect after Fontan palliation were excluded. The first day of aminophylline administration was considered to be day 0. Data were collected beginning 2 days prior to initial aminophylline dose and continuing through completion of the 5-day course of aminophylline. An intravascularly replete state was maintained in all patients; this was achieved by maintenance of blood pressure (invasive when available), central venous pressure, and heart rate within normative percentiles for age, by monitoring of physical exam signs such as capillary refill and mucosal moisture, and by obtaining serial laboratory markers including serum sodium and blood urea nitrogen concentrations.

AKI Criteria. For each day, we determined whether patients met criteria for AKI by Kidney Disease: Improving Global Outcomes (KDIGO) criteria.^25–27 Specifically, a patient was considered to have AKI on a specific date if he or she met 1 or more of the following criteria: 1) increase in serum creatinine by ≥ 0.3 mg/dL within 48 hours compared with baseline creatinine; 2) increase in serum creatinine to ≥1.5 times baseline creatinine; and 3) UOP volume <0.5 mL/kg/hr for 6 or more consecutive hours. Baseline serum creatinine was defined as the serum creatinine concentration on either 1 or 2 day(s), whichever was available, prior to the initial dose of aminophylline; preference was given to the value collected 2 days prior to aminophylline, if available.

Outcomes. The day following the last day during the study period that a patient met criteria for AKI (i.e., the first day on which none of the 3 criteria was met) and was not receiving either continuous renal replacement therapy (CRRT) or dialysis therapy was defined as the time of resolution. The primary outcome was AKI resolution during the study. Secondary outcomes included decreased serum creatinine concentration, decreased use of inotropes as assessed by VISCa (see below), and improved UOP.

Inotropic Support (VISCa). A daily vasoactive inotrope score (VIS) was also calculated for each patient receiving inotropes during the study period.^28,29 This VIS was further modified (VISCa) by expansion of the formula to include a continuous calcium infusion rate, which is frequently used for inotropic support in our institution,^30,31 particularly in the cardiovascular ICU:

VISCa = dopamine dose (μg/kg/min) + dobutamine dose (μg/kg/min) + 100 × epinephrine dose (μg/kg/min) + 10 × milrinone dose (μg/kg/min) + 10,000 × vasopressin dose (U/kg/min) + 100 × norepinephrine dose (μg/kg/min) + calcium chloride dose (mg/kg/hr).

Statistical Analysis. For statistical analysis, categorical data were summarized using percentages and frequencies. Continuous data were defined using median and the IQR. The Wilcoxon rank sum test and Fisher exact test were used to examine differences in the primary outcome for quantitative and categorical data, respectively. Cox proportional hazard regression models were used to assess bivariate associations between the outcome time to AKI resolution and several variables of interest including sex, race or ethnicity, age, body mass index, VISCa, and creatinine. All analyses were performed using SAS 9.4 (SAS Institute, Inc, Cary, NC).

Results

Among the 520 consecutive pediatric patients admitted to Le Bonheur Children's Hospital during our study period who had a diagnosis of AKI during their hospitalization, 29 received aminophylline in the pediatric or cardiovascular ICU. Seventeen of these patients received the standard AKI dosing of aminophylline, with 1 patient receiving standard AKI aminophylline therapy on 2 admissions and 1 patient receiving standard AKI aminophylline therapy for 2 separate treatment courses during 1 admission (separated by >5 days), yielding a total of 19 courses of therapy.

Descriptive Characteristics. Descriptive data for the study population among 19 therapies are provided in Table 1. During the study period, 4 patients were on dialysis or CRRT, 4 were on extracorporeal membrane oxygenation (ECMO), and 10 were on mechanical ventilation. There were no deaths during the study period, although 3 patients died prior to discharge. Of the 19 therapeutic courses of aminophylline, 12 (63%) resulted in resolution of AKI by the end of the study. The median age of patients with AKI that resolved during the study was 14 years, while the median age of patients with AKI that did not resolve during the study was 5.5 years. Table 2 summarizes the types of diuretics used by patients in each group.

Table 1.

Summary of Demographic and Clinical Characteristics

	All (N = 19)	AKI Not Resolved (n = 7)	AKI Resolved (n = 12)	p value
Age, median (IQR), yr	7.83 (1.2–15.4)	5.5 (0.3–8.0)	14 (3.0–15.6)	0.39
Sex, n (%)
Female	14 (74)	5 (71)	9 (75)	1
Male	5 (26)	2 (29)	3 (25)
Race, n (%)
Black	8 (42)	4 (57)	4 (33)	0.51
White non-Hispanic	5 (26)	2 (29)	3 (25)
Other or unknown	6 (32)	1 (14)	5 (42)
Weight, median (IQR), kg	20.9 (8.4–58.9)	12.7 (7–54.3)	48.9 (13.8–74.1)	0.21
BMI, median (IQR), kg/m2	20.5 (15.7–22.5)	17.3 (14.2–21.7)	21.4 (17.2–32.7)	0.13
LOS, median (IQR), days	42 (8–77)	58.4 (37.1–77.3)	25.7 (7.8–92.4)	0.39
Dialysis during study, n (%)	4 (21)	3 (43)	1 (8)	0.12
ECMO during study, n (%)	4 (21)	3 (43)	1 (8)	0.12
Vent during study, n (%)	10 (53)	5 (71)	5 (42)	0.35
Survival to discharge, n (%)	16 (84)	5 (71)	11 (92)	0.52

AKI, acute kidney injury; BMI, body mass index; ECMO, extracorporeal membrane oxygenation; LOS, length of stay; vent, mechanical ventilator

Table 2.

Summary of Diuretics by Group

	AKI Not Resolved, % (n) (n = 7)	AKI Resolved, % (n) (n = 12)
Bumetanide (continuous, IV)	57 (4)	33 (4)
Furosemide (intermittent, IV)	57 (4)	58 (7)
Chlorothiazide (intermittent, IV)	57 (4)	42 (5)
Combination loop + thiazide	43 (3)	33 (4)

AKI, acute kidney injury; IV, intravenous

Creatinine. The median (IQR) serum creatinine for the entire cohort decreased from 1.5 (0.7–1.9) mg/dL on the first day of aminophylline administration to 0.9 (0.35–1.6) mg/dL on the last day of the 5-day course (Table 3). Of note, following initiation of aminophylline, the median creatinine remained consistently higher among patients who met criteria for AKI resolution during the study than among the patients whose AKI did not resolve during the study (Figure 1).

Table 3.

Serum Creatinine Trend Between Groups

	Baseline	Day 0	Day 4
All, median (IQR), mg/dL	1.3 (0.50–1.70)	1.5 (0.70–1.90)	0.9 (0.35–1.60)
AKI resolved, median (IQR), mg/dL	1.35 (0.60–1.95)	1.6 (0.85–2.15)	1.1 (0.40–2.00)
AKI not resolved, median (IQR), mg/dL	1.1 (0.16–1.70)	1 (0.22–1.70)	0.55 (0.16–0.80)

AKI, acute kidney injury

Figure 1.

Median serum creatinine by day

Urine Output. In the group whose AKI resolved during the study, the median UOP increased 19% (p = 0.0063) during the first 24 hours following initiation of aminophylline before returning to approximately the pre-therapy value, where it remained relatively unchanged (Figure 2).

Figure 2.

Median UOP by day

Inotropic Support. Nearly half (9 out of 19) of the therapeutic courses did not include inotropic support during the study; 5 treatment courses from each group (5 with AKI resolution and 5 with AKI that did not resolve) received inotropes and/or continuous calcium chloride during the study. The median VISCa was unchanged throughout the study period for the group whose AKI did not resolve. In the group whose AKI did resolve, there was a decrease from a median VISCa of 7.38 on day 0 to a median of 3.85 on day 4.

Discussion

Our data demonstrate a trend toward improvement in serum creatinine and resolution of AKI with IV aminophylline administration, although none of the trends reached statistical significance likely due to small sample size. Additionally, our data suggest a marked diuretic effect early in the therapeutic course, with an increased UOP on the day following the initial aminophylline dose. We did observe a clinically significant improvement in median serum creatinine, even in the group that did not meet criteria for AKI resolution (Table 3).

We did not observe a statistically significant improvement in UOP from initial to final dose of aminophylline; however, as seen in Figure 2, we did observe a marked increase in median UOP on the day following the first dose of aminophylline in the patients whose AKI did resolve during the study; in patients whose AKI did not resolve, the UOP remained relatively stable during the study. This could represent a clinically significant effect of diuresis early in therapy, and it fits with the anecdotally observed increase in UOP 1 to 2 days following initiation of aminophylline therapy. A 2014 study by Axelrod et al¹³ found that aminophylline was useful to augment UOP in children with AKI in the cardiac ICU; however, the subsequent double-blinded, placebo controlled randomized control trial did not support aminophylline use in the post-cardiopulmonary bypass period due to lack of efficacy.²³

Our results also indicate a low rate of AKI resolution during our study period. Patients whose AKI did not resolve during our study period had a lower median serum creatinine than patients whose AKI resolved; thus, a relatively smaller increase in serum creatinine could have led to failure to resolve by the KDIGO criteria. Recent studies report both a positive effect on serum creatinine and/or UOP^{11,13,15,20,32} and a negative effect,²³ with a 2020 review article concluding that aminophylline is associated with a reduction in serum creatinine concentration without effect on incidence of AKI, UOP, or mortality.³³ Additionally, the marked age difference between the group whose AKI resolved and the group whose AKI did not resolve during the study (median 14 years and 5.5 years, respectively) could suggest that our definition of AKI resolution is not sufficiently sensitive across the pediatric age range, given lower normal creatinine values and higher UOP volume in younger patients.

Notably, there is no consensus definition of AKI in the pediatric population. The most commonly used criteria include the Risk, Injury, Failure, Loss, End Stage criteria's modification for use in the pediatric population, pediatric Risk, Injury, Failure, Loss, End Stage (pRIFLE), the Acute Kidney Injury Network (AKIN) criteria, and the KDIGO classification system. The KDIGO system is an attempt at integrating the pRIFLE and AKIN classification systems. Several studies have compared AKI incidence using the different diagnostic criteria; for example, Sutherland et al³ found incidences of 51.1%, 37.3%, and 40.3% by applying the pRIFLE, AKIN, and KDIGO criteria, respectively. A comparative study found that the pRIFLE system was the most sensitive test for AKI, the AKIN system was more specific, and KDIGO criteria fell between the 2 systems that it was designed to integrate. An extensive discussion of diagnostic criteria for pediatric AKI is beyond the scope of this manuscript, however. For our study, we chose to use the KDIGO criteria as the most recent evolution of the diagnostic criteria and because, based on nephrologist preference, KDIGO criteria are currently being used in other ongoing AKI studies at our institution.

Further, given the large proportion of patients who were on mechanical ventilation (53%), ECMO (21%), and CRRT or dialysis (16%), and a median hospital stay of 42 days, this cohort was severely ill and may not be representative of many of the patients who might have the greatest benefit from aminophylline in the pediatric critical care setting. Pediatric AKI literature has focused primarily on longer-term outcomes, such as length of stay and mortality, and few studies report duration or time to resolution in their population. One study reported a median time of renal recovery (defined as return to within 25% of baseline serum creatinine and UOP >1 mL/kg/hr for 24 hours) as 30 hours,³⁴ however the median length of hospital stay for their cohort was 8 days, suggesting that their population was less severely ill than ours. A recent study of the association between AKI duration and outcomes notes that persistent AKI (defined as AKI duration >48 hours) is associated with a higher illness severity and increased risk of morbidity and mortality.³⁵ Additionally, the authors concluded that increased AKI duration, regardless of degree of AKI severity, is associated with worse outcomes; they suggest that addition of AKI duration criteria to the definition of AKI may increase identification of high-risk AKI subpopulations.

We chose to analyze the VISCa as a surrogate marker of disease severity, hypothesizing that higher illness severity would be associated with higher VISCa values. However, less than half of our patients were on inotropes or continuous calcium chloride, leaving a very small population to assess. Notably, the VISCa in the group of patients whose AKI did not resolve during the study was low from the start of treatment. This could suggest that they were less severely ill or that they were undersupported hemodynamically on inotropes. Our unit practices prescribe low doses of inotropes even in markedly hypotensive and vasoplegic patients. This could contribute to inotropic undersupport, evidenced by lack of AKI resolution.

We did observe a clinically significant improvement in VISCa in the group whose AKI resolved (Figure 3), suggesting a correlation between improvement in disease severity and improvement in AKI. Our data suggest that, in patients provided appropriate hemodynamic support, use of aminophylline may be associated with an improvement in renal function. This is a plausible association, given the proposed mechanism of action of the drug of renal afferent arteriolar dilation would lead to increased renal blood flow. An alternative explanation is that, as illness severity improved, inotropes were able to be weaned due to improved hemodynamics, and end organ improvement resulted. However, in a retrospective study, it is difficult to propose causal relationships and this should be further explored in future studies.

Figure 3.

Median Vasoactive Inotrope Score + CaCl (VISCa) by day

We believe our use of the modified VISCa to be mathematically valid because we used the calculation uniformly on all patients, and the numerical weighting assigned to calcium chloride was similar to the mathematical results of the established VIS score.^28,29 This modification allows for appropriate mathematical inclusion of calcium chlo-ride drip rate as an inotrope^30,31 and enables comparison of inotropic requirements among patients.

There are some limitations to this study, including its sample size and single-institution retrospective nature. Therefore, our results may not be generalizable to a larger population or other institutions. As a retrospective study that relied on ICD-10²⁴ codes for identification of patients with AKI, an inherent limitation is the potential for missed and improperly identified patients with AKI. Record review did demonstrate AKI present in all patients who were included, however the possibility remains that there were additional patients with AKI whose coding failed to incorporate the appropriate ICD-10²⁴ code, and who therefore were not considered for this study.

Another limitation was the potential selection bias associated with a lack of standardization of aminophylline administration and patient laboratory monitoring. Criteria for using aminophylline were based on practitioner preference, leading to inconsistent timing of initiation. Additionally, there was a lack of standardization of the criteria used to diagnose AKI in our study participants. Currently, particularly in the cardiac ICU at our institution, diagnosis of AKI is sometimes based on cystatin C values, which are not part of widely accepted AKI diagnostic criteria. This lack of standardization could lead to inconsistent recognition of initial AKI diagnosis, which is an important variable in assessing time to resolution of AKI. In an attempt to maintain consistency in the values from our retrospective data set, baseline serum creatinine for our study was defined as the value obtained 1 to 2 days prior to initiating aminophylline; this introduces a possible source of error because this may not represent a true baseline value and would affect not only if or when patients met criteria for AKI but also if or when they met criteria for resolution of AKI. There were incomplete data sets, again, due to lack of standardization of patient laboratory monitoring and reporting of urine quantity, which may affect our results. The use of diuretics in our patients may have increased the UOP in our cohort, falsely elevating UOP in a patient who may otherwise have met criteria for oliguria. In the absence of a method for comparing diuretics of multiple mechanisms, routes, and doses, it is difficult to account for this, and many researchers simply disregard the UOP criteria of the AKI definition. Finally, as a descriptive study, it is impossible to attribute any of the observed effects to aminophylline, and future studies will be needed to address this.

Despite these limitations, we believe our data to be an accurate description of AKI resolution in critically ill pediatric patients treated with aminophylline at our institution. Additionally, these findings will be used to conduct a prospective randomized controlled trial comparing aminophylline to standard of care, which is currently in progress. Future directions could also include a prospective, placebo-controlled, multicenter randomized control trial.

Conclusion

In this single-institution, retrospective, descriptive study, aminophylline may be associated with resolution of AKI in 63% of patients during the time from initiation of aminophylline therapy to completion of the 5-day course. Additionally, we observed a trend toward decreased inotropic support, which may be extrapolated to assume decreased clinical severity. Another finding was a trend toward decreased serum creatinine during aminophylline administration. We also observed a spike in UOP on the day following administration of aminophylline therapy, suggesting an early diuretic effect from the therapy. Given the size and illness severity in our population, further studies are needed to determine the generalizability of our findings. The authors are aware of at least 1 ongoing clinical trial in this area, which will specifically compare the effect of preoperative aminophylline administration to saline placebo in the attenuation of renal injury in neonates and young infants following cardiac surgery.³⁶ Future studies will enable further characterization of the role of aminophylline in the treatment of AKI in critically ill pediatric patients.

ABBREVIATIONS

AKI

acute kidney injury

AKIN

Acute Kidney Injury Network

CRRT

continuous renal replacement therapy

ECMO

extracorporeal membrane oxygenation

ICU

intensive care unit

KDIGO

Kidney Disease: Improving Global Outcomes

pRIFLE

Pediatric Risk, Injury, Failure, Loss, End Stage

UOP

urine output

VIS

vasoactive inotrope score

VISCa

vasoactive inotrope score modified to include continuous calcium infusion