Incidence of acute kidney injury in dogs with systemic mycotic infections treated with amphotericin B (1996‐2020)

Abstract

Background

Amphotericin‐B (AmB) is an essential medication for the treatment of life‐threatening systemic mycoses but the incidence and risk factors for acute kidney injury (AKI) after its administration are not known in dogs.

Objective

Determine the incidence of and risk factors for AKI in dogs receiving AmB.

Animals

Fifty‐one client owned dogs receiving AmB for the treatment of systemic mycoses.

Methods

Retrospective study. Signalment, potential risk factors, AKI development (creatinine ≥0.3 mg/dL from baseline), drug formulation (deoxycholate [AmB‐D] or lipid complex [ABLC]), dose, and treatment duration were recorded. The probability of an AKI diagnosis was evaluated using a log‐rank test. The incidence of AKI and odds ratios were calculated for potential risk factors.

Results

Incidence of AKI was 5/12 (42%) for dogs receiving AmB‐D and 14/39 (36%) for dogs receiving ABLC. Of the 19 dogs that developed AKI, 16 (84%) continued treatment after a pause in the planned dosing protocol. Fifty percent of dogs received a cumulative dose of 6.9 mg/kg for AmB‐D and 22.5 mg/kg for ABLC (P < .01) at time of AKI diagnosis. ICU hospitalization (odds ratio [OR] 0.21, 95% confidence interval [CI]: 0.58‐0.87) and inpatient status (OR 0.25, 95% CI: 0.07‐0.86) were associated with decreased odds of AKI.

Conclusions and Clinical Importance

Incidence of AKI with AmB is common but does not always preclude continued treatment. The incidence of AKI is similar between AmB‐D and ABLC, but dogs receiving ABLC tolerated a higher cumulative total dose before AKI diagnosis.

Abbreviations

ABLC

amphotericin B lipid complex

AKI

acute kidney injury

AmB

amphotericin

AmB‐D

amphotericin deoxycholate

BCS

body condition score

CI

confidence interval

CKD

chronic kidney disease

IQR

interquartile range

NSAID

nonsteroidal anti‐inflammatory

OR

odds ratio

1. INTRODUCTION

Systemic mycoses, including histoplasmosis, blastomycosis, aspergillosis, cryptococcosis, and coccidioidomycosis, are rare but important infections of dogs. Dogs affected with systemic mycoses could present with disseminated disease that requires aggressive antifungal therapy. Amphotericin B (AmB), a polyene macrolide antibiotic, has been used in the treatment for invasive mycoses because of its rapid action and broad‐spectrum fungicidal activity. ¹ Amphotericin B has high affinity for ergosterol in fungal cell membranes, where it binds and creates a defect in the membrane. ¹ , ² It has poor oral bioavailability (0.2%‐0.9%) necessitating parenteral administration to reach therapeutic plasma concentrations. ³ Adverse effects associated with AmB parenteral administration include infusion reactions and more delayed effects, such as nephrotoxicosis. ⁴

There are several formulations of AmB, each with unique pharmacokinetic and toxicokinetic properties. Conventional AmB is complexed with sodium deoxycholate (AmB‐D; Fungizone, Xediton Pharmaceuticals Inc., Ontario, CA). It can cause cytotoxicosis to the distal renal tubules and potentially vasoconstriction that alters renal blood flow, leading to an increased risk of acute kidney injury (AKI). ⁵ , ⁶ , ⁷ To minimize nephrotoxicosis, AmB‐D is generally administered to dogs three times weekly, spaced at least 48 hours apart, as a 1 to 2 hour infusion with IV fluid diuresis. ⁸ , ⁹ Newer, lipid‐based amphotericin B formulations have a superior safety profile and have largely replaced the use of AmB‐D in people. ¹⁰ , ¹¹ Amphotericin B lipid complex formulation (ABLC; Abelcet, Lediant Biosciences, Gathersburg, MD) has fungicidal activity that is equal to that of AmB‐D and has reduced nephrotoxicosis in rats. ¹² This formulation allows for drug delivery in a ribbon‐like phospholipid suspension that increases tissue uptake of ABLC by the reticuloendothelial system. ¹³ ABLC is delivered primarily to sites of inflammation, resulting in reduced plasma concentrations and thereby, reduced nephrotoxicosis. ¹³ In people, there is a reduction in mean serum creatinine (Cr) in patients receiving ABLC, ¹¹ and ABLC does not significantly affect serum Cr in people with pre‐existing renal disease. ⁹ Other lipid‐based formulations such as liposomal amphotericin B (AmBisome, Astella Pharma, Northbrook, IL) have superior safety profiles to even ABLC; however, veterinary use is limited because of cost. ⁵ , ¹⁴

Limited data are available assessing the risk of kidney injury in clinical cohorts of dogs receiving AmB to treat systemic mycoses and factors that might increase the odds of developing an AKI. In one clinical trial that included 11 dogs with blastomycoses treated with ABLC, AKI was not reported. ¹⁵

The goal of our study was to determine the incidence of AKI in dogs with systemic mycoses treated with AmB and to assess differences in AKI incidence for dogs administered AmB‐D or ABLC. We hypothesized that AmB‐D would be associated with a greater incidence of AKI than ABLC. We hypothesized that dogs with pre‐existing kidney disease, recent anesthesia, or co‐administration of nephrotoxic drugs would have greater odds of developing an AKI.

2. METHODS

2.1. Animals and enrolment criteria

Medical records of dogs that received one or more doses of parenteral ABLC or AmB‐D prescribed at the University of California, Davis Veterinary Medical Teaching Hospital (VMTH) from 1996 to 2020 were reviewed. To be eligible for inclusion, dogs had to have serum Cr measured ≥2 times in the 2 weeks before the first AmB administration and at least every 4 days during the amphotericin course of treatment. Dogs were excluded if Cr increased by ≥0.3 mg/dL from baseline in the 2 weeks before AmB treatment or if a cause of kidney injury (ie, ureteral obstruction) was identified before AmB treatment. Azotemic dogs (Cr ≥1.4 mg/dL) were included as long as the azotemia did not increase in the 2 weeks before treatment.

Signalment and potential risk factors that were considered for diagnosis of an AKI were collected from the medical record. This included AmB formulation, cumulative total dose of AmB >20 mg/kg, starting dose of AmB, intensive care unit (ICU) hospitalization at start of treatment, inpatient status at start of treatment, administration of nonsteroidal anti‐inflammatory drugs (NSAIDs) within 2 weeks before or during AmB treatment, anesthesia within 2 weeks before or during AmB treatment, dehydration on physical examination at original presentation noted within the medical record, presence of neurologic deficits at any time point, and presence of chronic kidney disease (CKD) or fungal pyelonephritis at start of treatment. An arbitrary cut off using 20 mg/kg for the total dose of AmB was used as a potential risk factor as this is roughly 50% to 75% of the estimated dose routinely prescribed for AmB based on hospital protocol. ICU hospitalization indicated that dogs were specifically hospitalized in the ICU while inpatient status included all dogs in the ICU as well as dogs in intermediate and general hospital wards. Dogs hospitalized in ICU were considered to have highly dynamic clinical status in need of intensive serial monitoring requiring trained technical staff; ICU hospitalization also included any dog that required oxygen or ventilator support. Dogs hospitalized in the intermediate care ward needed more oversight than the general wards, such as an hourly respiratory watch, but were overall considered stable. Intravenous fluids and catheters were permitted in all hospital wards. Neurologic deficits were considered present if a dog's ability to drink and maintain hydration was impacted by neurologic disease. Chronic kidney disease was defined as repeated measurements of a stable Cr ≥1.4 mg/dL within the last 2 weeks before initiating treatment with 2 or more ultrasonographic changes consistent with CKD documented. Dogs could be considered to have chronic kidney disease and be non‐azotemic (Cr <1.4 mg/dL) if minimally concentrated urine (urine specific gravity <1.030) on a urinalysis before fluid therapy was present along with ultrasonographic changes consistent with CKD. These ultrasonsdographic changes included decreased corticomedullary distinction, reduced renal size, irregular renal contours, decreased size, pyelectasia, and mineral foci. ¹⁶ , ¹⁷ , ¹⁸ Fungal pyelonephritis was determined based on necropsy or inferred when ultrasonographic changes to the kidneys consistent with pyelonephritis (ie, renal papillary blunting, pyelectasia, and peri‐renal edema) were present along with a current positive urine fungal culture.

Body condition score (BCS) on a scale of 1 to 9, hydration status, weight, and Cr were recorded at each treatment date, at least 14 days before the first treatment, and up to 1 week after the final treatment. An ideal BCS was defined as a 4 or 5. BCS values of 1 through 3 were categorized as underconditioned while BCS 6 through 9 were categorized as overconditioned. Body condition score analysis was performed using the BCS at outset of therapy. The fungal diagnoses and diagnostics used to determine diagnosis were also recorded. The proportion of dog breeds in the study sample were compared with the proportion of dog breeds among all individual dogs presenting to the University of California, Davis VMTH from 1996 to 2020.

2.2. Amphotericin treatment

The institutional protocol for AmB was followed depending on formulation. For AmB‐D, a starting dose was typically administered at 0.5 to 1 mg/kg and generally escalated gradually up to 3 mg/kg 3 times a week for a total of 4 weeks but the protocol was not standardized in this study. For intravenous administration of AmB‐D, 0.9% NaCl was administered at 4 mL/kg over 1 hour before and after treatment. AmB‐D was reconstituted using sterile water to a dose of 5 mg/mL. The desired dose was further diluted with D5W to 0.5 mg/mL and administered over 4 to 6 hours. For subcutaneous administration of AmB‐D, the reconstituted amphotericin in sterile water was added to a 400 mL bag of 0.45% NaCl in 2.5% dextrose and administered subcutaneously.

Typically, 3 mg/kg of ABLC was administered 3 times a week for a total of 4 weeks but the infusion protocol was not standardized in this study. For intravenous administration of ABLC, the desired dose was diluted to 1 mg/mL in D5W. Before amphotericin infusion, 0.9% NaCl was administered intravenously at 4 mL/kg over 1 hour. The intravenous line was then flushed with 10 mL of 5% dextrose, and the desired dose of amphotericin was administered over 2 hours. After infusion, 10 mL of 5% dextrose was administered, and another 0.9% NaCl at 4 mL/kg over 1 hour was administered.

The formulation of amphotericin (AmB‐D or ABLC), dose per treatment (mg/kg), total cumulative dose (mg/kg), frequency of administration, route of administration (subcutaneous or intravenous), duration of treatment, number of treatments, and infusion‐related adverse events were recorded for each dog. For dogs that received multiple courses of AmB, only data from the first course of therapy was included. A course of therapy was defined as regular administration of a single AmB formulation administered on an every other day or at least 3 times a week basis with up to a week of temporarily discontinuing amphotericin if concerns of AKI arose. Creatinine and protocol of amphotericin at time of reinstitution for subsequent courses were clinician dependent. The number of additional courses of amphotericin was recorded but not further evaluated. Amphotericin was typically administered for 1 month but dogs could be continued on treatment for longer than a month as long as the dog was tolerating treatment and fungal disease was ongoing. Dogs were included in this study regardless of whether they completed the full intended course, and reasons for cessation of therapy were recorded. Survival to discharge, time to death, and time to euthanasia were noted.

2.3. AKI definition

The IRIS grading guidelines ¹⁹ were applied, defining an AKI as a sustained increase in Cr. A sustained increase was defined as 2 or more sequential increases in Cr ≥0.3 mg/dL as compared with baseline collected before the first AmB treatment.

2.4. Statistical analyses

Analyses were performed using commercial software (GraphPad Prism version 9.0f, San Diego, CA). Continuous variables were assessed for normality using the D'Agnostino and Pearson omnibus normality test. Normally distributed values are presented as a mean and standard deviation, and statistical analysis was performed by Student's t‐test. Data that were not normally distributed are presented as a median with interquartile range and analyzed with the Mann‐Whitney U test. Incidence of AKI was defined as the number of dogs developing an AKI over the study period. Chi‐squared analysis and Fisher's exact tests were performed to assess if proportions of dogs receiving either formulation differed for a given outcome. Odds ratios (OR) were used to assess risk factors for increased odds of developing AKI. Comparison of time and dose to event proportions for formulations were performed using a log‐rank (Mantel‐Cox) test. Confidence intervals (CI) were calculated with 95% confidence, and a significant P value was defined <.05.

3. RESULTS

3.1. Study sample

Between January 1, 1995 and January 2, 2020, 68 dogs were prescribed AmB at the VMTH. Fifty‐one dogs were included and 17 dogs were excluded (Figure 1).

FIGURE 1.

CONSORT flow diagram. AmB, amphotericin B. AKI, acute kidney injury.

Mean age of dogs in this study was 4.8 years (±2.6 years). Mean body weight was 26.0 kg (±12.9 kg). There were 25/51 (49%) spayed females, 4/51 (8%) intact females, 15/51 (29%) castrated males, and 7/51 (14%) intact males. The majority of dogs in this study were purebred (39/51, 76%). These included 8 Labrador retrievers, 7 German shepherd dogs, 2 Staffordshire terriers, and 2 Pugs. Labrador retrievers and German shepherd dogs were over‐represented relative to the general population of dogs that presented during this time frame. Fungal diagnoses included coccidioidomycosis (14), aspergillosis (9), cryptococcosis (8), hyalohyphomycosis (5), Rasamsonia piperina (3), blastomycosis (3), phaeohyphomycosis (2), histoplasmosis (2), Penicillium spp. (1), Talaromyces helices (1), and unidentified (3). Median duration of AmB treatment was 12 days (IQR 7‐17, range 1‐117), median number of treatments was 5 (IQR 2‐8, range 1‐43), and median total dose was 15 mg/kg (IQR 10‐20, range 0.5‐65).

Of the 51 dogs, 12 (24%) received AmB‐D and 39 (76%) dogs received ABLC. Six of the 51 dogs received subsequent courses of therapy but only data from the first treatment course is included. Characteristics including age, weight, BCS, and duration of clinical signs before starting administration of amphotericin were similar across both groups (Table 1).

TABLE 1.

Demographic Information and treatment information for dogs receiving amphotericin B deoxycholate (AmB‐D) and amphotericin B lipid complex (ABLC).

	AmB‐D (n = 12)	ABLC (n = 39)
Age (years)	5.3 (1.6‐9.1)	4.7 (0.5‐10.9)
Weight (kg)	26.0 (1.9‐65)	23.1 (3.4‐44.3)
BCS	4.7 (2.0‐7.0)	4.3 (1.0‐8.0)
Duration of clinical signs pretreatment (days)	48 (4‐139)	50 (7‐301)
ICU hospitalization	4/12 (33%)	14/39 (35%)
CKD	1/12 (8%)	3/39 (8%)
Incidence of AKI	5/12 (41%)	14/39 (36%)
Inpatient	7/12 (58%)	22/39 (56%)
Survival to discharge	11/12 (91%)	26/39 (67%)

Note: Numerical data is presented as a mean and range. Proportions are presented as a ratio and percentage. No statistically significant differences were present between dogs receiving AmB‐D and ABLC.

Abbreviations: ABLC, amphotericin lipid complex; AKI, acute kidney injury; AmB‐D, amphotericin deoxycholate; BCS, body condition score; CKD, chronic kidney disease; ICU, intensive care unit.

Four dogs had evidence of CKD before treatment was initiated. Two dogs had a stable baseline azotemia (baseline Cr 1.5 and 1.8 mg/dL) while the other 2 were non‐azotemic (Cr <1.4 mg/dL) but had ultrasonographic evidence of CKD and minimally concentrated urine. Three of the 4 dogs with CKD received ABLC and 1 received AmB‐D. Sixteen dogs had fungal pyelonephritis of which 13 received ABLC and 3 received AmB‐D.

3.2. Incidence of AKI

Nineteen of the 51 enrolled dogs (37%) dogs treated with AmB were diagnosed with an AKI. The median magnitude of change in Cr at time of first AKI was 0.4 mg/dL (IQR 0.3‐0.7, range 0.3‐4.1; Figure 2). Fifteen of the 19 (79%) dogs had an IRIS grade I AKI, 3/19 (16%) had an IRIS grade II AKI, and 1/19 (5%) had an IRIS grade III AKI. Renal replacement therapy was not performed in any of the dogs. Three of the 4 dogs with CKD were diagnosed with an AKI. Five out of 12 (42%) of dogs treated with AmB‐D and 14/39 (36%) dogs treated with ABLC group were diagnosed with an AKI, respectively (Table 2). There was no difference in the incidence of AKI between formulations (P = .70).

FIGURE 2.

Magnitude of change in creatinine at time of first AKI. Each dot represents an individual dog. The bar represents the median change in creatinine. AKI, acute kidney injury.

TABLE 2.

Summary of subgroups.

	All dogs	Dogs with AKI	AmB‐D	AmB‐D with AKI	ABLC	ABLC with AKI
Number of dogs	51	19	12	5	39	14
Median cumulative dose (mg/kg) to AKI diagnosis		7		4.5*		8.5*
Median total dose (mg/kg)	9	14.5	1.5*	6	10*	15
Median number of treatments to AKI diagnosis		5		9*		4*
Median total number of treatments	5	10	3	10	5	11
Median days to AKI diagnosis		13		22*		10.5*
Median days of total treatment	12	28	4	28	8	28.5

Abbreviations: AmB, amphotericin B; ABLC, amphotericin lipid complex; AKI, acute kidney injury; AmB‐D, amphotericin deoxycholate.

^* P < .05 when comparing outcomes for dogs receiving either AmB‐D or ABLC.

The median total dose of AmB was 14.5 mg/kg (range 3‐56.5 mg/kg) for dogs that were diagnosed with an AKI and 7 mg/kg (range, 0.5‐65 mg/kg) for dogs that were not diagnosed with an AKI (P = .07). At 23 days, there was a 50% probability of dogs receiving AmB‐D being diagnosed with an AKI verses 32 days for dogs receiving ABLC (Figure 3, P = .90). Furthermore, at a cumulative dose of 6.9 mg/kg, there was a 50% probability that dogs receiving AmB‐D would be diagnosed with an AKI, and for dogs receiving ABLC, there was a 50% probability of diagnosis of AKI at cumulative dose of 22.5 mg/kg (Figure 4; P < .01). However, amongst only the dogs that were diagnosed with an AKI, the median time to its diagnosis was 22 days for AmB‐D and 10.5 days for ABLC (P = .04). Amongst the dogs that were diagnosed with an AKI, the median cumulative dose at time of AKI diagnosis was 4.5 mg/kg for AmB‐D and 8.5 mg/kg for ABLC (P = .03, Table 2).

FIGURE 3.

AKI‐free survival based on cumulative time comparing dogs receiving AmB‐D vs ABLC using log‐rank proportion comparison (P = .90). X‐axis represents days of treatment. Y‐axis represents probability of not developing an AKI during treatment. AmB, amphotericin B; ABLC, amphotericin lipid complex; AKI, acute kidney injury; AmB‐D, amphotericin deoxycholate.

FIGURE 4.

AKI‐free survival based on cumulative dose comparing dogs receiving AmB‐D vs ABLC using a log‐rank proportion comparison (P < .01). X‐axis represents dose over time. Y‐axis represents probability of not developing an AKI during treatment. AmB, amphotericin B; ABLC, amphotericin lipid complex; AKI, acute kidney injury; AmB‐D, amphotericin deoxycholate.

3.3. AmB dosing strategy

Of the 12 dogs receiving AmB‐D, 8 (67%) received IV administration while 4 (33%) received SQ administration. Three of the 8 dogs (38%) receiving IV administration of AmB‐D and 2/4 (50%) of the dogs receiving SQ administration of AmB‐D were diagnosed with an AKI. No significant difference was observed comparing incidence of AKI by route of administration (P = 1.00).

3.4. Outcomes

Of the 19 dogs diagnosed with an AKI, 16 (84%) continued to receive AmB after a modification of the intended dosing protocol. Modification of the dosing protocol was clinician dependent and included a temporary halt in AmB administration. Timing of restarting therapy and any change in dose or frequency of AmB administration were left to the discretion of the clinician. Of those 16 dogs, 10 were diagnosed with a second AKI. Nine of these 10 dogs continued to receive AmB and were diagnosed with a third AKI. The 3 dogs that received a diagnosis of AKI and did not continue to receive AmB included 1 dog that had therapy halted due to concerns for ongoing nephrotoxicosis, 1 dog that was euthanized due to progressive fungal disease, and 1 dog that was diagnosed with an AKI after completion of the intended course of therapy.

Creatinine at the start of amphotericin treatment was significantly lower than Cr at the end of the first course of amphotericin (P < .01). The median increase in Cr was 0.1 mg/dL (IQR −0.1 to 0.4) by the end of amphotericin treatment.

Thirteen dogs were euthanized and 1 died before completing the intended full 4 weeks of AmB. One dog was euthanized due to concerns of nephrotoxicosis secondary to AmB. Of the 13 dogs, 1 (8%) was euthanized due to financial constraints, and 12 (92%) were euthanized due to complications (including nephrotoxicosis) or poor quality of life secondary to their systemic mycoses. Eleven out of 12 (92%) of dogs receiving AmB‐D and 26/39 (67%) of dogs receiving ABLC were alive at the end of their course of treatment (P = .10).

3.5. Non‐AmB risk factors for development of AKI

Potential risk factors associated with the diagnosis of AKI were evaluated. Factors evaluated included dehydration at presentation (OR 0.95, 95% CI: 0.24‐3.8, P = .94, n = 12), pre‐existing CKD (OR 5.8, 95% CI: 0.56‐60.48, P = .14, n = 4), NSAID administration within 2 weeks before or during amphotericin treatment (OR 3.15, 95% CI: 0.83‐11.96, P = .09, n = 12), anesthesia within 2 weeks before or during amphotericin treatment (OR 1.2, 95% CI: 0.38‐3.86, P = .74, n = 20), ICU hospitalization at the start of treatment (OR 0.21, 95% CI: 0.58‐0.87, P = .03, n = 18), inpatient hospitalization at the start of treatment (OR 0.25, 95% CI: 0.07‐0.86, P = .02, n = 34), fungal pyelonephritis (OR 1.06, 95% CI: 0.3‐3.36, P = .91, n = 16), and signs of neurologic disease (OR 1.21, 95% CI: 0.39‐3.81, P = .74, n = 29; Figure 5).

FIGURE 5.

Forest plot depicting odds ratios with 95% CI for various potential risk factors independent of formulation. Dots represent odds ratio and bars represent 95% confidence intervals. CKD, chronic kidney disease; ICU, intensive care unit; NSAID, nonsteroidal anti‐inflammatory.

Body condition score was recorded for 45/51 (88%) dogs. Twenty‐five of 45 dogs (56%) had an appropriate body condition, while 12/45 (26%) were underconditioned and 8/45 (18%) were overconditioned. Nine dogs with ideal body weight, 3 dogs that were underconditioned, and 5 dogs that were overconditioned were diagnosed with an AKI. Body condition was not found to be associated with risk of diagnosis of AKI.

3.6. AmB‐associated risk factors for development of AKI

AmB‐D dogs had an odds ratio of 1.28 (CI 0.34‐4.78) as compared with dogs with ABLC for developing AKI (P = .72). Dogs who received <20 mg/kg in total of AmB had no significant increase in odds of AKI as compared with dogs who received ≥20 mg/kg (OR 0.48, 95% CI: 0.14‐1.68, P = .11).

The median starting dose of AmB‐D was 0.5 mg/kg (IQR 0.5‐1.3) vs 2 mg/kg (IQR 1.0‐3.0) for ABLC. There was not a significant difference in incidence of AKI between dogs receiving 3 mg/kg of ABLC as a starting dose compared with dogs receiving lower starting doses (≤2 mg/kg, P = .76).

4. DISCUSSION

In our study, the overall incidence of AKI in dogs treated with AMB was 37%, and incidence was similar between the 2 commonly administered formulations (AmB‐D [42%] and ABLC [36%]). However, dogs that were diagnosed with AKI received a higher cumulative dose before onset of AKI if they were administered ABLC vs AmB‐D. The incidence of AKI was similar to human studies, which have reported incidences of 37 and 45%. ²⁰ , ²¹ Our study supports that nephrotoxicosis is a risk with AmB administration, regardless of formulation administered. Clinicians should be vigilant in monitoring for development of AKI, including development of non‐azotemic AKI, especially when dogs receive repeated doses of AmB.

Development of an AKI rarely precluded continued treatment with AmB in this cohort with only 16% of dogs having treatment halted upon diagnosis of an AKI. Reason for cessation of therapy was often due to completion of the prescribed course of AmB or worsening of or complications from fungal disease. In a minority, nephrotoxicosis was considered a reason for cessation of therapy but the exact aspect of the nephrotoxicosis that led to cessation (ie, magnitude of change in Cr, absolute Cr, clinical condition) could not be inferred from the medical record due to the retrospective nature of this study. Of the 84% of dogs that continued to receive amphotericin, over 50% had repeat AKI. Continued administration of AmB in the face of AKI highlights the morbidity and risk of death associated with disseminated fungal diseases. While nephrotoxicosis is concerning, the risk of an AKI could be a tolerable outcome in light of close monitoring of Cr and adjustment in treatment protocol. This discrepancy further demonstrates that AKI was not a factor leading to cessation of therapy in the majority of cases. This study did not assess if azotemic dogs returned to baseline but complete clinicopathologic recovery from an AKI occurs in 36% of human patients who developed AKI after administration of liposomal AmB. ²²

Dogs receiving ABLC achieved a higher median cumulative dose before an AKI onset as compared with AmB‐D. However, direct comparison of the dose of ABLC and AmB‐D is not feasible due to formulation difference and potential bias. Of the subset of dogs that developed an AKI, dogs that there were receiving ABLC were more likely to have an AKI earlier in their relative course of treatment as compared with AmB‐D. We suspect this finding could be due to higher starting doses of ABLC as compared with AmB‐D in light of the experimentally reduced risk of nephrotoxicosis with ABLC. ⁴ Clinicians might perceive ABLC as safer than AmB‐D, which could translate to prescription of higher doses of ABLC as compared with AmB‐D. Dosing protocol was not standardized in this study. While increased nephrotoxicity of AmB‐D is on potential explanation as to why dogs with ABLC could withstand a total dose and longer treatment duration, potential bias in which dog received each formulation needs to be taken into account. AmB‐D has historically been available at a reduced cost as compared with ABLC, ²³ and it is possible that dogs receiving AmB‐D were not as closely monitored in hospital due to owner cost constraints. Although this study spanned a 24‐year period, there was minimal change in frequency of prescription of AmB formulation with ABLC being generally preferred over the entire timeframe.

Potential risk factors beyond AmB formulation and dose associated with an AKI were assessed. Inpatient and ICU hospitalization at the outset of treatment were associated with decreased odds of AKI diagnosis. We hypothesize that inpatient and ICU hospitalization allowed for closer monitoring of dogs during treatment. No other risk factors were identified that were significantly associated with the diagnosis of an AKI in our study. However, with the low sample size, this study might not have been powered to detect some risk factors for development of AKI that have been identified in people. CKD and NSAID use have not been associated with increased odds of AKI in humans receiving liposomal based AmB but have been associated with an increased odds of AKI in humans receiving conventional AmB. ²⁰ , ²⁴ In people, AKI risk increases if liposomal amphotericin was not prescribed on a lean body weight. ²⁵ While BCS was not associated with risk of AKI in this study, it is recommended that AmB be prescribed based on lean body weight. Fungal pyelonephritis was not identified as a risk factor in this study but the possibility of fungal die off contributing to AKI after amphotericin administration cannot be excluded.

In people there are variable results regarding whether AKI risk is dose‐dependent or related to total cumulative dose, ²⁰ , ²⁶ and no such studies have been performed in veterinary medicine to date in clinical animals. This study did not find an increased odds of AKI in dogs receiving higher cumulative doses of AmB, by comparing dogs who received <20 mg/kg AmB total against dogs who received ≥20 mg/kg.

There were several limitations of this study. While Cr was monitored in all cases as a part of the inclusion criteria, the frequency of serum Cr measurements was variable for different dogs. Serum Cr concentrations were measured in 48 to 120 hour intervals, and it is possible that short‐term fluctuations were missed. Another limitation of this study was the limited sample size, particularly dogs receiving AmB‐D. This limitation is likely due to a shift in clinicians' preference for liposomal‐based formulations due to concerns for AmB‐D induced nephrotoxicosis. Our study was designed to assess differences in rates of AKI between AmB formulations, and not powered to assess for additional risk factors; therefore, further investigation into additional risk factors should be pursued. Our study is likely underpowered to evaluate these risk factors, as illustrated by the wide confidence intervals for some of the potential risk factors. Assessment of risk of AKI with different formulations amongst dogs with CKD was not possible due the low number of dogs with CKD but comparison of incidence of AKI with different AmB formulations in this study sample would be clinically relevant and should be an area of future study. Because of the retrospective design, in‐hospital care, including adjunctive therapy, fluid diuresis, and dosing regimens, was not standardized.

In conclusion, diagnosis of an AKI in association with AmB administration occurred with an incidence of 37% in our study sample, but nephrotoxicosis might not preclude further administration of AmB with close monitoring and changes in dosing protocols. Dogs with ABLC received a greater cumulative dose before onset of AKI as compared with dogs receiving AmB‐D. ICU hospitalization and inpatient status at outset of treatment were protective against diagnosis of an AKI.

CONFLICT OF INTEREST DECLARATION

Authors declare no conflict of interest.

OFF‐LABEL ANTIMICROBIAL DECLARATION

Authors declare no off‐label use of antimicrobials.

INSTITUTIONAL ANIMAL CARE AND USE COMMITTEE (IACUC) OR OTHER APPROVAL DECLARATION

Authors declare no IACUC or other approval was needed.

HUMAN ETHICS APPROVAL DECLARATION

Authors declare human ethics approval was not needed for this study.

Chan JC, Dear J, Palm C, Reagan K. Incidence of acute kidney injury in dogs with systemic mycotic infections treated with amphotericin B (1996‐2020). J Vet Intern Med. 2023;37(3):1030‐1037. doi: 10.1111/jvim.16728