Introduction

The nephrotoxic effects of plants from the Lilium and Hemerocallis species have been recognized and documented in cats for over 20 years.^1–6 The development of clinical signs such as vomiting and hypersalivation may occur within several hours following exposure, and signs associated with acute kidney injury (AKI) typically develop within 24 hours.^1,7–9 In cats where treatment is delayed, mortality rates as high as 50% to 100% have been reported.^1,3 Unfortunately, there is little known about the specific toxin resulting in kidney injury beyond that it is a water-soluble compound found in all parts of the lily plant.⁹ This knowledge gap has hindered the development of evidenced-based therapeutic recommendations for lily toxicosis.

The current standard of care for lily ingestion includes gastrointestinal decontamination in the acute exposure period to minimize toxin absorption followed by IV fluid administration for 48 hours.¹⁰ Baseline and daily assessment of kidney function (eg, BUN, creatinine, and phosphorous) and urinalysis for urine specific gravity, glucosuria, and cylinduria reflecting tubular injury are recommended to evaluate for evidence of AKI.¹⁰ These treatment recommendations are largely based on a small descriptive observational study⁵ of 25 cats with known or suspected lily ingestion. In that study, 23 of 25 cats were treated with IV fluid therapy in addition to other supportive measures; only 2 of the hospitalized cats (9%) developed an AKI, and all cats survived. The true effect of supportive measures, including gastrointestinal decontamination and IV fluid therapy, on the prevention of AKI in these cats is unknown and would ultimately require a randomized controlled trial. Considering the potentially fatal consequences of lily exposure in cats, it is unlikely that such a trial would be performed.

Despite the current recommendations for IV fluid therapy following acute lily exposure, not all cat owners are able to pursue this treatment for various reasons, including financial limitations, lack of access to 24-hour veterinary facilities, or other cat-specific factors. In these cats, following initial gastrointestinal decontamination, outpatient (OP) management with daily SC fluids and monitoring of kidney function over 48 hours may be recommended as an alternative treatment strategy. To our knowledge, the prevalence of AKI in cats managed with this alternative treatment strategy has not been investigated. Therefore, the purpose of our study was to compare the outcomes and prevalence of AKI in cats with known or suspected lily exposures that were treated in the inpatient (IP) group with IV fluids or the OP group. A secondary aim of our study was to investigate for factors associated with the development of AKI. We hypothesized that cats in the OP group would be associated with higher prevalence of AKI and worse outcome as compared to cats that were managed in the IP group.

Methods

Medical records at the Matthew J. Ryan Veterinary Hospital at the University of Pennsylvania were searched for cats presenting for lily exposure between January 2011 and January 2024. Medical records were examined to determine whether cats met the inclusion criteria. Cats were included in the study if the suspected or known lily exposure occurred within 48 hours of hospital presentation and if the following criteria were met: (1) time of presentation from lily exposure was documented, (2) creatinine at the time of initial hospital presentation was available, and (3) at least 2 different time points evaluating creatinine in both the hospitalized and OP group cats were available. Known lily exposure was defined as cats that had witnessed ingestion, vomited pieces of lily, chewed on the flower or leaf, and/or had pollen on their body. Suspected lily exposure was defined as cats that were in the room with lilies. Patients were excluded if the time of lily exposure was > 48 hours prior to hospital presentation, if they were previously hospitalized for lily exposure at another facility prior to referral, if the lily ingested was not nephrotoxic (eg, peace lily), if no baseline creatinine was documented, if there were < 2 time points in which creatinine was assessed, if the medical records were incomplete, or if other comorbidities unrelated to lily exposures were present. Primary care veterinarians were contacted for records documenting recheck creatinine and outcome if cats did not return to our hospital for recheck.

Information retrieved and analyzed from the medical records included date of hospital presentation, age, sex, breed, intact or altered sex, whether the lily exposure was known or suspected, lily species, and the maximum time from lily exposure to presentation. Initial decontamination treatments were recorded, including whether emesis was induced, whether emesis was successful and if lilies were seen in the vomitus, whether gastric lavage was performed, and whether activated charcoal was administered and dose. Finally, cats were categorized into either the hospitalized IP group or the OP group. In the cats that were in the IP group, the duration of hospitalization and maximum rate (mL/kg/h) of IV fluid administration were recorded. In the OP group, the dose (mL/kg) of SC fluids was recorded if administered. For both groups of cats, creatinine values and AKI grade at initial presentation (baseline), 0 to 24 hours, 24 to 48 hours, 48 hours to 2 weeks, whether the cat developed an AKI at any point (yes/no), whether AKI grade was static or improved when comparing the last documented to baseline AKI grade (yes/no), and outcome (alive or dead/euthanized) were recorded. Acute kidney injury grade was defined by the International Renal Interest Society (IRIS) guidelines for diagnosis and grading of AKI.¹¹ Cats were classified as having at least a stage 1 AKI if creatinine was above 1.6 mg/dL at baseline or if at any point they had an increase in creatinine ≥ 0.3 mg/dL. Static AKI grade was classified as unchanged AKI grade comparing the last documented AKI grade to baseline AKI grade.

Results

One hundred seventy-nine cats that were presented to the hospital for lily exposure were initially identified. After the inclusion and exclusion criteria were applied, 67 cats were excluded from analysis due to the following reasons: lack of follow-up creatinine values (38 cats); exposure occurring for > 48 hours (14 cats); missing or incomplete medical records (8 cats); lack of baseline creatinine values (3 cats); other comorbidities defined as an iatrogenic uroabdomen, pancreatitis, and suspected triaditis (1 cat each); and exposure to a non-nephrotoxic lily (1 cat). One hundred twelve cats met the inclusion criteria: 96 cats (85.7%) in the IP group and 16 cats (14.3%) in the OP group.

The median age of all cats was 3.4 years (range, 0.25 to 19.5 years). One hundred eight cats (96.4%) were mixed-breed cats, of which 102 (91%) were domestic shorthair and 6 (5.4%) were domestic longhair; 2 cats (1.8%) were Bengal, 1 cat (0.9%) was a Persian, and 1 cat (0.9%) was a Siamese. Sixty-one cats (54.4%) were neutered males, 48 cats (42.9%) were neutered females, 2 cats (1.8%) were intact males, and 1 cat (0.9%) was an intact female. Neither age nor sex was significantly different between the IP and OP groups (P = .2 and P = 1.00 respectively). The type of lily was known for 41 cats (36.6%): 11 (26.8%) were daylilies, 10 (24.4%) were stargazer lilies, 8 (19.5%) were Asiatic lilies, 8 (19.5%) were tiger lilies, and 4 (9.8%) were Oriental lilies. The lily type was unknown in 71 cats (63.4%). February and May (16 cats each), followed by June (12 cats), were the months with the greatest number of cats with known or suspected lily exposures. The least number of exposures occurred in January and November (5 cats each); however, lily exposures occurred every month during the year.

The maximum time from lily exposure to presentation was a median of 3.75 hours (range, 0.25 to 48 hours) in the IP group and 9 hours (range, 0.25 to 24 hours) in the OP group. This was not significantly different between the 2 groups (P = 0.2). In the IP group, 83 cats (86.5%) had known lily exposure, and 13 cats (13.5%) had suspected exposure. In the OP group, 9 cats (56%) had known lily exposures and 7 cats (43.7%) had suspected lily exposure. There were significantly more cats with known exposures in the IP group compared to the OP group (P = .02).

Emesis was attempted in 35 cats (36.5%) in the IP group and 2 cats (12.5%) in the OP group. Of the cats where emesis was induced, success of emesis was noted in 23 cats (65.7%) in the IP group and 1 cat (50%) of the OP group. Of the cats that had successful emesis, only 10 cats (43.5%) in the IP group and the 1 cat (100%) in the OP group had parts of the lily recovered. Activated charcoal was administered to 33 cats (34.4%) in the IP group and 2 cats (12.5%) in the OP group. The median dose of activated charcoal was 3 mL/kg (range, 1 to 4 mL/kg) in the IP group and 3.9 mL/kg (range, 2.9 to 4.9 mL/kg) in the OP group. There was no difference in emesis attempted, success of emesis, lily seen upon emesis, or activated charcoal administration in cats treated in the IP versus the OP group (Table 1; P > .05 for all comparisons). In the OP group, 13 cats (81.3%) received SC fluids, with the median SC fluid dose of 15 mL/kg (range, 9.9 to 31.6 mL/kg). In the IP group, the median duration of hospitalization was 50 hours (range, 15 to 92 hours). The median maximum IV fluid rate administered in the IP group was 2 mL/kg/h (range, 4 to 6 mL/kg/h).

The median admission baseline creatinine in both the IP and OP group was 1.3 mg/dL (IP range, 0.6 to 2.5 mg/dL; OP range, 1 to 2.1 mg/dL; P = .8). There were 45 cats (46.9%) in the IP group and 7 cats (43.8%) in the OP group that had a documented AKI on serial bloodwork at any time point. Of these cats that had AKIs, 27 (60%) in the IP group and 4 (57.1%) in the OP group had a static or improved AKI grade when the last documented AKI grade was compared to baseline AKI grade. In the IP group, 51 cats (53.1%) did not have an AKI, 17 cats (17.7%) had a maximum AKI grade of I, 27 cats (28.1%) had a maximum AKI grade of II, and 1 cat (1.1%) had a maximum AKI grade of III. No cats in the IP group had a maximum AKI grade of IV or higher. In the OP group, 9 cats (56.2%) did not have an AKI, 3 cats (18.8%) had a maximum AKI grade of I, and 4 cats (25%) had a maximum AKI grade of II. No cats in the OP group had a maximum AKI grade of III or higher (Table 2). In the OP group, 1 cat developed an AKI at the first 24-hour time point (creatinine increased from 1.3 mg/dL at baseline to 2 mg/dL at 0 to 24 hours). This cat was subsequently hospitalized and treated with IV fluids, though still considered to be part of the OP group. Serial creatinine measurements in this cat showed improved creatinine concentrations throughout the hospitalization period from 1.5 mg/dL (creatinine at 24 to 48 hours) to 1.3 mg/dL at the final time point (creatinine at 48 hours to 2 weeks). No association of treatment group (IP vs OP group) and development of AKI at any time point was identified (P = .6). Additionally, age, sex, known versus suspected exposure, maximum time to exposure, whether emesis was attempted, the success of emesis, whether lilies were recovered during emesis induction, activated charcoal administration, SC fluid administration, and duration of hospitalization were not associated with an AKI at any point (P > 0.05).

There was a significant association between treatment group (IP vs OP group) and outcome, with 96 cats (100%) surviving in the IP group and 14 cats (87.5%) surviving in the OP group (P = .02). For the 2 nonsurvivor cats in the OP group, both cats were suspected lily exposures and were euthanized with IV pentobarbital. The reason for euthanasia was not documented in 1 cat, but this cat was noted to be lethargic, anorexic, vomiting, and adipsic for 2 consecutive visits despite SC fluids prior to euthanasia. The cat presented approximately 24 hours after suspected lily exposure, was assessed to be clinically euhydrated on physical examination, and had no prior bloodwork documenting preexisting renal disease. The cat was considered to have a renal azotemia at presentation based on a baseline admission creatinine of 2.1 mg/dL (grade II AKI) and a baseline urinalysis documenting a urine specific gravity of 1.024 with no sediment abnormalities. The cat had a subsequent creatinine of 1.8 mg/dL (grade II AKI) at the 0- to 24-hour time point. The cat was ultimately euthanized the next day, and a creatinine measurement was not performed. The second euthanized OP cat had reported vomiting, lethargy, depressed mentation, and progressive azotemia despite SC fluids. Euthanasia was reported to be secondary to financial constraints. This cat had a baseline admission creatinine of 1.1 mg/dL, a 0- to 24-hour creatinine of 1.6 mg/dL (grade I AKI), and a 24- to 48-hour creatinine of 1.8 mg/dL on the day of euthanasia (grade II AKI).

Discussion

In this study, we did not identify an increased prevalence of AKI in the OP group (43.8%) as compared to the IP group (46.9%; P = .6). The lack of an association between treatment groups and the development of AKI are likely due to the higher overall prevalence of AKI identified in both IP and OP groups of cats. This is in comparison to a previous study⁵ reporting a 9% prevalence of AKI in lily-exposed cats treated with IV fluid therapy. The higher prevalence of AKI detected in our study in both groups could be due to the larger number of cats included (112 cats). Additionally, the application of the 2016 IRIS AKI grading guidelines also allowed for the detection of nonazotemic AKI patients, which was not available at the time of the previous study.⁵ Despite the higher prevalence of AKI identified in both populations of cats, many of the cats with an AKI had a static or improved AKI grade when the last documented AKI grade was compared to the baseline AKI grade (60% in the IP group and 57.1% in the OP group). Only 1 cat in our study had an AKI grade of III, which subsequently improved to an AKI grade of II. No cats had a grade IV or V AKI (creatinine > 5 mg/dL). This is encouraging given that there are previous reports documenting persistent azotemia, oligo-anuric kidney failure requiring renal replacement therapy, and death in cats following lily exposure.^1,3,4,8,9 Our results could have suggested that the prevalence of severe kidney injury is low.

Despite the higher prevalence of AKIs detected in our study, all cats in the IP group survived. The same survival rate was previously reported in a smaller study⁵ of 25 lily-exposed cats from the same institution. In our study, only cats with acute lily exposure of 48 hours or less and without hospitalization for the same incident at another facility were included in the study. This led to the exclusion of 14 cats, which could have influenced the 100% survival rate in the IP group. The decision to exclude cats with lily exposures of > 48 hours was made for the following reasons: (1) to evaluate differences between treatment in the IP and OP groups on the prevention of AKI due to the previous association of AKI and poor outcome with delayed treatment,^1,3,7,8 (2) because specific treatment may not be recommended for cats without evidence of AKI at > 48 hours, and (3) to eliminate previously hospitalized cats with an established AKI that were referred to our hospital, as their treatment may have differed from what is provided at our institution. This last group of cats could potentially represent failure of an IP protocol if they were treated with IV fluids. There is also a possibility that excluded cats that presented > 48 hours after lily exposure could have had an AKI; this population of cats warrants further investigation.

The standard of care at our hospital for lily exposure is hospitalization with IV fluid therapy, which likely limited the number of cats in the OP group. Every cat in the IP group received IV fluids, which could have contributed to a favorable outcome by minimizing renal ischemia during hypoperfused states, diluting and enhancing elimination of the lily nephrotoxin, and prevention of tubular obstruction from casts.¹² In the OP group, 14 of the 16 cats received SC fluids. This may have had a similar effect in preventing AKI, as the prevalences in both groups were not significantly different. However, this finding should be interpreted cautiously, as our study was ultimately underpowered to detect a treatment effect on the prevalence of AKI. Future studies are needed to identify the water-soluble compound in lilies, to investigate the pharmacokinetics of the toxin such as toxic dose and the effect of IV or SC fluids on toxin excretion, and to investigate prevention and survivability following kidney injury.

The significantly lower survival rate identified in the OP group (87.5% vs 100%) can be attributed to the euthanasia of 2 cats. Both euthanized cats had a grade II AKI (maximum creatinine of 2.1 mg/dL and 1.8 mg/dL respectively) and persistent clinical signs of lethargy, anorexia, and vomiting. It is possible that these 2 cats could have survived with hospitalization and additional supportive care, as this was observed in a different OP cat (described above) that was subsequently hospitalized following a documented grade II AKI at 0 to 24 hours. Despite a higher mortality of cats treated in the OP group, the prognosis was still favorable; however, it is important to note that there were significantly more cats with known exposures in the IP group than the OP group (P = .02), with only 9 of the 16 cats in the OP group having known exposure to lilies. One possible explanation for this difference could have been that cat owners in the OP group chose OP treatment over hospitalization due to a lower index of suspicion for true lily exposure, although other factors such as finances could have also contributed to the decision.

As a secondary aim of our study, we attempted to identify factors that may be associated with the development of an AKI. Of the factors evaluated (age, sex, known versus suspected exposure, maximum time to exposure, SC fluid administration, and duration of hospitalization), none were found to be associated with the development of an AKI. Interestingly, despite a low number of cats receiving gastric decontamination with emesis induction (36.5% in the IP group and 12.5% in the OP group) and activated charcoal administration (34.4% in the IP group and 12.5% in the OP group), this was also not associated with the development of an AKI. These findings could have suggested that gastric decontamination with emesis induction and activated charcoal administration may have little effect on the prevalence of AKI and outcome. Previous studies^13–16 investigating the efficacy of gastric decontamination for toxin removal in dogs showed variable rates of gastric recovery ranging from 17% to 62% depending on timing of emesis. To our knowledge, the efficacy of gastric toxin removal through emesis in cats has not been established. However, a study¹⁷ of foreign body removal in cats found a 50% success of emesis for some or all of the foreign objects. A low rate of toxin removal could have contributed to the lack of correlation between gastric decontamination and the prevalence of AKI and outcome. However, because our study was observational and treatments were not standardized among patients, additional controlled prospective studies are needed to further evaluate the effect of different treatments on the development of AKI.

This study had several additional limitations. The IRIS AKI grading criteria may have led to an overdiagnosis of clinically insignificant AKIs, as 17 cats in the IP group and 3 cats in the OP group were diagnosed with a grade I AKI. The use of an absolute increase in serum creatinine concentration > 0.3 mg/dL from baseline might have resulted in these cats being falsely classified as having an AKI. Serum creatinine concentrations can have daily variability in addition to being affected by hydration status, fasted versus fed states, and muscle mass. In a study¹⁸ of healthy cats, day-to-day variability in creatinine concentration identified a median variation of 0.3 mg/dL. Additionally, there can be variabilities in sample-to-sample blood analyzer measurements in creatinine that could have affected creatinine measurements and led to an under- or overestimation of creatinine and thereby affected the AKI grade. Creatinine measurements beyond 2 weeks after lily exposure were also not available to assess for ongoing renal recovery in cats with persistent AKI. Lastly, urinalysis at admission or at the recheck creatinine measurement time points was not routinely available to evaluate urine specific gravity or for the presence of urinary casts.

Our study also included cats from a tertiary referral hospital where many patients may be referred for hospitalization, which limited the number of cats in the OP group. As a result, cats in this study may not accurately reflect the population of cats seen in the general practice and nonreferral settings. Additionally, cats with both suspected and known lily exposures were included in our study as treatment recommendations, for these cats are typically similar. The inclusion of cats with suspected lily exposure could have led to an underestimation in true prevalence of AKI. Finally, given the retrospective observational design of our study, cats were not randomized to treatment protocols. As such, cat-, client-, and clinician-related factors may have influenced treatment groups and outcome.

Overall, our study found a higher prevalence of mild to moderate AKI in both groups of cats as compared to previous reports of acute lily-exposed cats. Despite this higher prevalence of AKI, the overall survival rate remained excellent. Although this study should not be used to make a definitive recommendation on the ideal treatment approach (IV fluids or SC fluids), our results were encouraging in that no cats in the OP group were diagnosed with a moderate AKI (stage III or higher) or required renal replacement therapy. Larger controlled prospective studies are needed to further confirm these findings and establish evidence-based treatment recommendations.

Disclosures

The authors have nothing to disclose. No AI-assisted technologies were used in the generation of this manuscript.

Funding

The authors have nothing to disclose.

ORCID

E. Reineke https://orcid.org/0000-0001-9911-8698