Pilot study of side effects and serum and urine concentrations of amoxicillin–clavulanic acid in azotemic and non-azotemic cats

Abstract

Objectives

The aims of this study were to determine the side effect frequency and serum and urine drug concentrations of amoxicillin–clavulanic acid in cats with and without azotemic chronic kidney disease (azCKD).

Methods

Owners whose cats had been prescribed amoxicillin–clavulanic acid completed a survey regarding the occurrence and type of side effects, and whether treatment was altered as a result. Cats were defined as azCKD (serum creatinine concentration >2.0 mg/dl, urine specific gravity [USG] <1.035 with a clinical diagnosis of chronic kidney disease) and without azCKD (serum creatinine concentration <2.0 mg/dl). Data were assessed with Fisher’s exact test. Serum and urine samples were obtained from client-owned cats with azCKD (n = 6) and without azCKD (n = 6, serum creatinine concentration <1.8 mg/dl, USG >1.035) that were receiving amoxicillin–clavulanic acid. Amoxicillin and clavulanic acid were measured with liquid chromatography coupled to tandem mass spectrometry and compared between groups with a Mann–Whitney test. Correlation between serum creatinine and drug concentrations in urine and serum was determined using Spearman’s rank test.

Results

Sixty-one surveys were returned (11 azCKD cats and 50 without azCKD cats). No significant difference in the presence of side effects or type of side effects was seen between groups; however, significantly more azCKD cats had more than one side effect (P = 0.02). More owners of azCKD cats reported that an alteration in treatment plan was necessitated by side effects (55% vs 12%; P = 0.008). Urine amoxicillin was significantly lower in cats with azCKD (P = 0.01) and serum amoxicillin trended toward significance (P = 0.07). Serum amoxicillin concentration was positively correlated with serum creatinine (P = 0.02; r = 0.62) and urine amoxicillin concentration was negatively correlated with serum creatinine (P = 0.01; r = –0.65).

Conclusions and relevance

The data suggest that cats with azCKD have altered pharmacokinetics of amoxicillin, which may contribute to an increased incidence of multiple side effects.

Introduction

Chronic kidney disease (CKD) is a common disease of older cats, affecting up to 30–40% of cats over 10 years of age. ¹ Antimicrobials are frequently prescribed in this patient population due to the high prevalence of urinary tract infections (UTIs), which is reported to be approximately 20–30% of CKD cats. ² Therapy with a penicillin (amoxicillin) or potentiated penicillin (amoxicillin–clavulanic acid) are often prescribed as both drugs provide sufficient activity against the most common bacterial pathogens in the feline lower urinary tract; Clavamox (Zoetis Animal Health) (amoxicillin–clavulanic acid) is approved by the US Food and Drug Administration for use in feline UTIs. ³

In human patients with decreased renal function, potentiated penicillins have decreased renal clearance, resulting in higher serum concentrations in comparison with patients with normal renal function.^4,5 Furthermore, drug exposure increases in relation to the severity of renal impairment. The mean clearance for amoxicillin can be reduced by as much as 90% and clavulanic acid up to 70% in patients with varying degrees of renal impairment. ⁴ Potentiated penicillins are predominately excreted by the kidneys; therefore, measurement of urine concentrations of the drug has been shown to have a linear association with creatinine clearance in humans. With decreasing creatinine clearance, a subsequent decrease in urinary excretion of the drug is observed. ⁵ It has thus been recommended in human patients with reduced kidney function that the dose of potentiated penicillins be altered (ie, decreased dose or decreased frequency) based on creatinine clearance to avoid excessive drug accumulation.

From an anecdotal perspective, cats with CKD tend to have an increased frequency and severity of side effects when administered amoxicillin–clavulanic acid than cats with normal kidney function. This may reduce client compliance to complete treatment recommendations, particularly when treatment duration is prolonged. Based on the information available for humans with renal disease, we hypothesized that cats with CKD may be more prone to side effects with amoxicillin–clavulanic acid due to decreased renal clearance and drug accumulation. The purpose of this pilot study was therefore two-fold: (1) to determine if cats with azotemic CKD (azCKD) have an increased frequency of side effects from amoxicillin–clavulanic acid in comparison with cats without azCKD using a client survey format; and (2) to determine if cats with azCKD have altered serum and urine concentrations of amoxicillin–clavulanic acid in comparison with cats without azCKD. This information will provide important data for future considerations of dose alterations of potentiated penicillins in cats with azCKD to ultimately improve quality of life and client compliance, and decrease unwanted side effects.

Materials and methods

Client survey

Owners whose cats had been prescribed amoxicillin–clavulanic acid at Colorado State University and The Ohio State University were given a survey, either at the time of receiving their prescription or they were emailed a survey after completion of their cat’s antibiotic treatment. Antibiotic treatment was prescribed by the attending clinician and not for the purposes of the study. The survey contained questions regarding the occurrence of side effects (yes/no), specific type of side effect (vomiting, diarrhea, decreased appetite, none) and whether treatment was altered as a result (see Appendix 1 in the supplementary material).Signalment, clinicopathologic and dosing data were obtained from the medical record for cats for which completed surveys were returned. Cats included in the azCKD group had a serum creatinine concentration >2.0 mg/dl (reference interval 0.8–2.4 mg/dl), urine specific gravity (USG) <1.035 and a clinical designation of CKD based on a previous history of CKD diagnosis, persistent renal azotemia and ultrasonographic changes consistent with CKD. Cats included in the non-azotemic group were those with a serum creatinine concentration <2.0 mg/ dl and with no previous history of clinical indication of CKD or clinical signs of CKD. A serum creatinine cut off of 2.0 mg/dl was chosen to define cohorts in order to recruit cats with CKD in later International Renal Interest Society (IRIS) stages (late stage 2 and higher). This rationale is based on human studies documenting that impaired clearance of amoxicillin–clavulanic acid is more pronounced in humans with a greater degree of renal impairment. ⁴

Amoxicillin–clavulanic acid urine and serum concentrations

Eight client-owned cats with azCKD and seven client-owned cats without azCKD prescribed amoxicillin–clavulanic acid for reasons determined by their attending clinician were recruited. Cats were screened with a complete physical examination (including muscle mass score), complete blood count, serum biochemistry, urinalysis and total thyroxine to ensure no comorbidities existed that could interfere with drug absorption or metabolism (hepatic impairment, uncontrolled hyperthyroidism, gastrointestinal disease). Cats included in the azCKD group had a serum creatinine concentration >2.0 mg/dl, USG <1.035 and a clinical designation of CKD based on a previous history of CKD diagnosis, persistent renal azotemia and ultrasonographic changes consistent with CKD. Cats included in the no azCKD group were those with a serum creatinine concentration <1.8 mg/dl, USG >1.035 and no previous history of clinical signs or clinical diagnosis of CKD.Cats were prescribed amoxicillin–clavulanic acid (tablet or suspension) at a dose and frequency determined by the attending clinician. Based on a half-life of approximately 1.3 h of amoxicillin and 0.6 h of clavulanic acid in cats, the steady state of amoxicillin–clavulanic acid (based on amoxicillin) is estimated to occur by 10 h; therefore, cats returned after 5–7 days of therapy for collection of blood and urine samples.^6,7 Jugular or medial saphenous venipuncture was performed for blood collection and cystocentesis was performed for urine collection 1–2 h after a dose of amoxicillin–clavulanic acid, which correlates with peak serum concentrations.^6,7 Serum and urine samples were kept frozen at ‒80°C until analysis.The study was approved by the Institutional Animal Care and Use Committee and the Clinical Review Board at Colorado State University. Written informed consent was obtained from the owner or legal custodian of all cats described in this study for the procedures undertaken.

Amoxicillin and clavulanic acid analysis

Amoxicillin and clavulanic acid were measured simultaneously in the cat’s serum by liquid chromatography coupled to tandem mass spectrometry (LC/MS/MS) using an assay developed and validated in the University of Colorado Cancer Center Pharmacology Shared Resource located at Colorado State University. The instrumentation used was an Applied Biosystems 3200 Q-TRAP triple quadrupole mass spectrometer coupled to an Agilent 1200 high-performance liquid chromatography (HPLC) system and an HTC-PAL autosampler.

Unknown and quality control (QC) samples were prepared by adding known amounts of amoxicillin–clavulanic acid (0.05–400 µg/ml) into blank cat serum. Samples (unknowns, standards and QCs) were prepared for analysis by mixing a 50 µl serum sample with 5 µl ampicillin (100 µg/ml, internal standard) and adding 100 µl acetonitrile followed by vortex mixing for 5 mins to precipitate proteins. Samples were then centrifuged for 15 mins at 13,300 g, 100 µl of the resulting supernatant collected, transferred to a HPLC vial and mixed with 900 µl 2 mM ammonium acetate. Sample injection volume was 30 µl (using a 20 µl loading loop) and chromatography was carried out using a Waters Sunfire C8 5 µm column (4.6 × 50 mm) and a solvent system consisting of acetonitrile (solvent A) and 2 mM ammonium acetate (solvent B). Amoxicillin, clavulanic acid and ampicillin (internal standard) were eluted using a gradient starting at 2 solvent A:98 solvent B for the first 1.0 mins and then transitioning linearly to 90 solvent A:10 solvent B over a 1 min period and holding until 4 mins when the original 2 solvent A:98 solvent B was re-established over a 0.3 min period and held for the remainder of the 6.5 min total run time.The mass spectrometer was operated in negative ion mode utilizing a spray voltage of ‒4500 V and a source temperature of 550^°C. Multiple reaction monitoring (MRM) analysis was carried out for amoxicillin by monitoring ion transitions of 364.1 m/z → 129.0 m/z and 364.1 m/z → 206.0 m/z, for clavulanic acid by monitoring ion transitions of 197.8 m/z → 108.0 m/z and 197.8 m/z → 136.0 m/z and for ampicillin ion transitions of 348.0 m/z → 304.0 m/z. MRM conditions were optimized using internal algorithms and both Q1 and Q2 were operated at unit resolution.Assay performance was monitored using QC samples at three levels (0.075, 0.75 and 7.5 µg/ ml) and for both amoxicillin and clavulanic acid 100% (n = 24/24) of QC samples had accuracies >85%. The overall accuracy and precision (% coefficient variation) were 94.0% ± 4.5% for amoxicillin and 93.5% ± 3.9% for clavulanic acid, with a lower limit of quantitation of 0.05 µg/ml used for this analysis.

Statistical analysis

Data were analyzed with Prism software (GraphPad Software). Based on assessment for normality with D’Agostino and Pearson’s test, non-parametric statistical analyses were chosen. Fisher’s exact test was used to assess the distribution of the presence of side effects (yes/no), prevalence of specific side effects and changes in therapy between the azCKD and without azCKD groups collected from the survey data. For comparison of serum and urine amoxicillin–clavulanic acid concentrations between the azCKD and no azCKD groups, a Mann–Whitney test was used. Spearman’s rank correlation was used to identify a correlation between serum creatinine and amoxicillin–clavulanic acid concentrations between the two groups. For all analyses, a P value of <0.05 was considered to be statistically significant.

Results

Client survey

Seventy-three surveys were returned, with 12 being excluded due to incomplete data or the serum creatinine levels being unavailable. Eleven cats were categorized as azCKD (IRIS stages represented were stage 2 [n = 9], stage 3 [n = 1] and stage 4 [n = 1]), and 50 cats were categorized as not having azCKD. Signalment, creatinine and side effect results for each group are presented in Table 1.Overall, 6/11 (55%) owners of cats with azCKD and 20/50 (40%) owners of cats without azCKD reported side effects following amoxicillin–clavulanic acid administration. The difference in distribution of side effects (yes/no) and distribution of a specific type of side effect was not statistically significant between groups; however, significantly more azCKD cats had more than one side effect reported (P = 0.02) (Table 1). Additionally, more owners of azCKD cats reported that an alteration in treatment plan was necessitated by side effects, with 55% indicating that the antibiotic was stopped, changed or supportive care was added due to side effects, while only 14% of owners of cats without azCKD reported an alteration in treatment plan (P = 0.008).From a review of the medical records, additional medications were prescribed along with the amoxicillin–clavulanic acid prescription in the majority of cats in both groups. Of the azCKD cats, 81% were prescribed additional medication, with buprenorphine (n = 5) being the most common, followed by maropitant (n = 4), and one cat each receiving potassium, amlodipine, mirtazapine and subcutaneous fluids. Of the cats without azCKD, 80% were prescribed additional medications, with buprenorphine (n = 14) being the most common, followed by prednisolone (n = 6), gabapentin (n = 4), mirtazapine (n = 3), glargine (n = 2) and one cat each receiving cefovecin, robenacoxib, ondansetron, ursodiol, meloxicam, palladia, cyanocobalamin and famciclovir.

Table 1.

Summary of cats receiving amoxicillin–clavulanic acid and owner questionnaire survey results

	Azotemic CKD (n = 11)*	Without azotemic CKD (n = 50) †
Signalment
Sex
MC	4	17
FS	7	33
Age (years)	14 (4–20)	11 (1–18)
Creatinine (mg/dl)	2.4 (2.0–6.0)	1.4 (0.7–1.9)
Dose (mg/kg)	13.7 (8.9–20.2)	14.0 (12.7–17.6)
Side effects	6 (55)	20 (40)
Vomiting	5 (45)	11 (22)
Diarrhea	2 (18)	10 (20)
Decreased appetite	3 (27)	7 (14)
>1 side effect	5 (45) ‡	6 (12)

Data are n (%) or median (range)

^*Azotemic chronic kidney disease (CKD) defined as serum creatinine concentration >2.0 mg/dl, urine specific gravity <1.035 and a clinical designation of CKD

^†Without azotemic CKD defined as serum creatinine concentration <2.0 mg/dl and with no previous history of clinical indication of CKD or clinical signs of CKD

^‡P = 0.009

MC = male castrated; FS = female spayed

Amoxicillin–clavulanic acid concentrations

Fifteen client-owned cats with and without azCKD were considered for enrollment. One cat in the azCKD group was excluded after its azotemia resolved following dental extractions and normal renal ultrasound. One young cat in the no azCKD group was excluded owing to minimally concentrated urine (USG 1.031). One cat from the azCKD group was excluded owing to the degree of renal disease, which was considered mild, based on serum creatinine concentration (1.2 mg/dl), USG (1.029) and symmetric dimethylarginine (SDMA; 15 µg/dl). Six cats with naturally occurring azCKD and six cats without azCKD were eventually included in the analysis.The azCKD group consisted of three castrated males and three spayed females, with a median age of 15.5 years (range 4–20 years) and breeds included five domestic shorthairs and one Manx. Median serum creatinine concentration was 2.4 mg/dl (range 1.8–6.0 mg/dl) with the following IRIS stages represented: stage 2 (n = 5) and stage 4 (n = 1). All azCKD cats had evidence of muscle wasting on physical examination that varied from mild (n = 3) to moderate (n = 1) and severe (n = 2). The cat with a serum creatinine concentration of 1.8 mg/dl had historically had a serum creatinine concentration >2.0 mg/dl and had notable muscle atrophy on examination. Additionally, one cat with a serum creatinine concentration of 2.7 mg/dl had severe muscle wasting, which may have influenced the creatinine level. All azCKD cats had minimally concentrated to isosthenuric urine with a median USG of 1.015 (range 1.009–1.019). Concurrent conditions included controlled hyperthyroidism (n = 1), controlled diabetes mellitus (n = 1) and feline immunodeficiency virus (n = 1). The reason for antibiotic prescription included UTI (n = 2), pyelonephritis (n = 1), anal gland abscess (n = 1) and dermal abscess (n = 1). The median dosage of amoxicillin–clavulanic acid was 14 mg/kg (range 12.7–15.2 mg/kg) q12h PO. Three of six cats experienced side effects, including decreased appetite (n = 1), diarrhea (n = 1) and both decreased appetite and diarrhea (n = 1).

The group of cats without azCKD consisted of three castrated males and three spayed females, with a median age of 9 years (range 4–12) and breeds included five domestic shorthairs and one Himalayan. The median serum creatinine concentration was 1.0 mg/dl (range 0.7–1.8 mg/dl) and all cats had adequately concentrated urine with a median USG of 1.062 (range 1.054–1.073). The cat with a serum creatinine concentration of 1.8 mg/dl was suspected to have a mild pre-renal azotemia, as evident by the presence of hemoconcentration (hematocrit 53%). SDMA would have been useful for further characterization; however, it was not performed at the time. Only one cat had mild muscle atrophy on physical examination. One cat had concurrent diabetes mellitus, with the remainder of the cats having no additional comorbidities. The reasons for antibiotic prescription included dermal/bite abscesses (n = 3), followed by post-dental extractions (n = 1), UTI (n = 1; cat with diabetes mellitus) and lower urinary tract signs (n = 1). The median dose of amoxicillin–clavulanic acid was 13.1 mg/ kg (range 12–13.5 mg/kg). Four of six cats had side effects that included decreased appetite (n = 2), vomiting (n = 1) and diarrhea (n = 1).

Serum amoxicillin concentrations were higher in the azCKD group, trending toward significance (P = 0.07), with a median amoxicillin concentration of 17.3 µg/ml (range 2.3–21.4 µg/ml) compared with a median amoxicillin concentration of 5.6 µg/ml (range 1.1–10.0 µg/ml) in the group without azCKD (Figure 1). A moderate positive correlation between serum amoxicillin concentration and serum creatinine was also observed (P = 0.02; r = 0.62) (Figure 2). Urine amoxicillin concentrations were significantly lower in the azCKD group, with a median concentration of 50.5 µg/ml (range 8.1–158.5 µg/ ml), in comparison with the group without azCKD, which had a median urine amoxicillin concentration of 328.3 ng/ml (range 33.1–576.5 µg/ml) (P = 0.01) (Figure 3). In addition, there was a moderate negative correlation between urine amoxicillin concentrations and serum creatinine (P = 0.01; r = –0.65) (Figure 4).There was no difference in clavulanic acid serum or urine concentrations between groups. However, the azCKD group trended toward higher serum concentrations of clavulanic acid, with a median concentration of 15.0 µg/ml (range 0.5–29.2 µg/ml), than the group without azCKD (median concentration 4.3 µg/ml; range 0.05–13.5 µg/ml). There was no significant difference in urine clavulanic acid concentrations, with a median concentration of 82.5 µg/ml (range 7.0–247.0 µg/ml) in the azCKD group and a median concentration of 56.7 µg/ml (range 22.0–664.5 µg/ml) in the group without azCKD.

Figure 1.

Comparison of serum amoxicillin concentrations in cats with and without azotemic chronic kidney disease (azCKD). Median amoxicillin–clavulanic acid dosage in cats with azCKD was 14.0 mg/kg q12h (range 12.7–15.2 mg/kg q12h) and median dosage in cats without azCKD was 13.1 mg/kg q12h (range 12.9–13.5 mg/kg q12h). Serum amoxicillin concentrations were higher in the CKD cats, trending toward significance (P = 0.07). Bar represents median

Figure 2.

Serum amoxicillin concentration had a moderate positive correlation with serum creatinine when assessed in cats with a range of serum creatinine concentrations. Median amoxicillin–clavulanic acid dosage in cats with azotemic chronic kidney disease (azCKD) was 14.0 mg/kg q12h (range 12.7–15.2 mg/kg q12h) and median dosage in cats without azCKD was 13.1 mg/kg q12h (range 12.9–13.5 mg/kg q12h)

Figure 3.

Comparison of urine amoxicillin concentrations in cats with and without azotemic chronic kidney disease (azCKD). Median amoxicillin–clavulanic acid dosage in cats with azCKD was 14.0 mg/kg q12h (range 12.7–15.2 mg/kg q12h) and median dosage in cats without azCKD was 13.1 mg/kg q12h (range 12.9–13.5 mg/kg q12h). Urine amoxicillin was significantly lower in CKD cats (P = 0.01). Bar represents median

Figure 4.

Urine amoxicillin concentration was moderately negatively correlated with serum creatinine when assessed in cats with a range of serum creatinine concentrations. Median amoxicillin–clavulanic acid dosage in cats with azotemic chronic kidney disease (azCKD) was 14.0 mg/kg q12h (range 12.7–15.2 mg/kg q12h) and median dosage in cats without azCKD was 13.1 mg/kg q12h (range 12.9–13.5 mg/kg q12h)

Discussion

The results of this pilot study suggest that the pharmacokinetics of amoxicillin are altered in cats with azCKD that are administered amoxicillin–clavulanic acid. Similar to humans with renal impairment, the serum amoxicillin concentration was higher and the urine amoxicillin concentration was lower in the azCKD cat group and each was positively and negatively correlated with creatinine, respectively. The difference in serum amoxicillin concentrations did not reach significance, likely because of a limitation of sample size and a limited hospital population of IRIS stage 3–4 azCKD cats. It is predicted that this could be a repeatable finding in a larger population of patients and when evaluating cats with more advanced CKD.

Although the urine amoxicillin concentrations were significantly lower in the azCKD cats, the lowest urinary amoxicillin concentration measured (8.1 µg/ml) would still exceed the current feline Clinical Laboratory Standards Institute (CLSI) plasma drug concentration-based resistance breakpoint of 1 µg/ml of amoxicillin used for Escherichia coli. ⁸ This suggests that, despite altered pharmacokinetics, there is sufficient drug excreted into the urine, which correlates with clinical improvement. The cat with the lowest urine amoxicillin concentration was previously classified as stage 3, based on IRIS guidelines, and was treated for pyelonephritis after its creatinine concentration increased to 6.0 mg/dl. Cats with more advanced CKD (stage 4 or acute kidney injury) may have a more profound decrease in urinary concentrations and below minimum inhibitory concentration breakpoints, which could potentially lead to treatment failure. Cats without azCKD demonstrated significantly higher urine amoxicillin concentrations compared with azCKD cats, further suggesting that the evaluation of less conservative amoxicillin breakpoints for urinary isolates may be appropriate for cats, as the Clinical Laboratory Standards Institute (CLSI) has previously approved less stringent breakpoints for canine urinary isolates. ⁸

There was no statistical difference in serum or urine concentrations of clavulanic acid between groups; however, the azCKD group trended toward higher serum levels than the group without azCKD. Our small sample size is likely a contributing factor for the lack of statistical significance; however, there are also other complex factors that should be considered. Variable oral absorption of clavulanic acid has been documented in humans, cats and dogs, and there are also non-renal mechanisms of clavulanic acid clearance.^7,9–11 Multiple human studies have demonstrated clavulanic acid excretion to be less affected in renal impairment after a single intravenous bolus, as well as more rapid clavulanic acid excretion in anephric patients, suggesting the presence of a non-renal mechanism of clearance.^9,12 In addition, the mechanism of renal elimination for clavulanic acid is excretion by glomerular filtration, whereas amoxicillin is eliminated by renal tubular excretion, which may also contribute to variability in serum and urine concentrations in patients with primarily tubular disease.^4,13

The survey format had a limited ability to determine if the increase in number of side effects in the azCKD cats is a reflection of drug accumulation (of either amoxicillin or clavulanic acid), potential comorbidities or exacerbation of comorbidities, bias of owner reporting in more affected patients or the potential effect of other prescribed medications. For example, four CKD cats were prescribed maropitant, and this could have helped prevent side effects (as, indeed, none of these four cats experienced vomiting but did experience other side effects), so it is possible that side effects in CKD cats were underappreciated. Future studies should include additional survey questions to determine if side effects were attributable to amoxicillin–clavulanic acid rather than a sequalae of underlying disease or effect of other medications, and a standardized scoring system, including fecal scoring system, implemented prospectively during amoxicillin–clavulanic therapy. This would also provide a better foundation for assessing the relationship of side effects and serum concentrations.

Additional limitations of this study include no standardized amoxicillin–clavulanic acid dose or formulation (tablet vs suspension) and lack of reliable categorization of the various CKD stages or evaluation of glomerular filtration rate. Most of the azCKD cats had some evidence of muscle atrophy, which likely influenced their creatinine levels. Evaluation of SDMA may have been beneficial to provide additional information to characterize the degree of renal disease.

Conclusions

Serum and urine amoxicillin concentrations had a positive and negative correlation with serum creatinine, respectively, suggesting that altered pharmacokinetics exist in cats with azCKD and may be more apparent as disease severity progresses. Clavulanic acid disposition may be less affected by renal disease; however, its metabolism is more complex than amoxicillin and a conclusion could not be drawn from this study. In addition, altered drug metabolism may be an underlying etiology for an increased incidence of side effects in cats with azCKD receiving amoxicillin–clavulanic acid. Prospective studies with a larger population are required to determine if dose adjustments should be considered for azotemic feline patients.