Toxicity associated with ingestion of a polyacrylic acid hydrogel dog pad

Introduction

Polyacrylic acid is extremely hydrophilic and can absorb over 100 times its weight in water, producing a “hydrogel.” ² Hydrogels are polymeric materials that swell and retain a significant fraction (>20%) of water within their structure when they are placed in water. ²¹ This absorbent property has led to widespread commercial applications in medicine, engineering, environmental remediation, and the pharmaceutical industry. ²¹ Polyacrylic acid is also used in diapers, sanitary napkins, hand sanitizers, mascara, hair-styling products, and moisturizers. ²¹

One advantage of these superabsorbent polymers is that they are relatively nontoxic, with acute oral median lethal doses (LD₅₀) often >5 g/kg. ²¹ An acute oral LD₅₀ of >40 g/kg has been reported for Rhotex GS, a 15:85 mixture of sodium polyacrylic acid and water in rats. ³ Hydrogels also have relatively low toxicity following repeated exposure. For example, rats given a structurally related cross-linked polyacrylic acid polymer ≤5% in their diet for 32 d or ≤3 g/kg/d in their diet for 93 d did not exhibit signs of overt toxicity.^5,9

Although these polymers are considered nontoxic, significant adverse effects can be seen following massive ingestion. In 2015, a 3-y-old, 22.3 kg, castrated male Australian Shepherd mix was referred by a local veterinarian to the North Carolina State University Veterinary Hospital (NCSU-VH) for evaluation of acute onset of clinical signs (ataxia, vomiting, anxiety, intention tremors, and tachypnea) occurring within 24 h after the ingestion of ~0.35 kg piece of sodium polyacrylate–based hydrogel found within a commercial dog pad (estimated exposure = 15.7 g/kg). The owner reported that the outer plastic shell of the pet pad was torn open by the dog and the interior hydrogel material was missing. Visual examination of a vomitus sample collected shortly after ingestion verified the presence of the hydrogel material, thus confirming that exposure occurred. A complete blood count and serum biochemical analysis performed by the referring veterinarian showed azotemia and slightly elevated liver enzyme activities (Table 1). Initial neurologic examination at NCSU-VH revealed the presence of inappropriate mentation characterized by agitation and increased sensitivity to sound, profound ataxia and hypermetria in all 4 limbs, wide-based posture, intention tremors, bilateral absence of the menace response, and diffuse discomfort on palpation of the thoracolumbar spine. Based on the neurologic examination, the most likely neuroanatomic diagnosis was a multifocal process involving the forebrain and cerebellum. Although toxicity was strongly suspected, other differential diagnoses were considered, including infectious meningoencephalitis and neoplasia.

OPEN IN VIEWER

Table 1.

Summary of abnormal biochemical results observed in a dog after ingestion of a pet pad hydrogel.

Analyte	Unit	Value			NCSU-VH reference interval
		1 d after exposure	3 d after exposure	4 d after exposure
Urea	mmol/L	5.6	3.7	0.8	1.2–4.5
Creatinine	µmol/L	274	71	35	44–159
ALP	U/L	215	156	145	23–212
ALT	U/L	131	189	185	10–125
CK	U/L	NA	25,200	4,975	43–234

ALP = alkaline phosphatase; ALT = alanine aminotransferase; CK = creatine kinase; NA = not available; NCSU-VH = North Carolina State University Veterinary Hospital.

Laboratory tests for this dog explored potential toxic, infectious, and neoplastic etiologies. An ethylene glycol exposure test was negative. Magnetic resonance imaging (MRI) studies did not reveal any morphologic changes in the central nervous system (CNS), although focal ill-defined areas of contrast enhancement were noted in the left splenius, rectus capitis, and longus capitis muscles. These muscle lesions were felt to be consistent with inflammation secondary to muscle spasm. Cerebrospinal fluid (CSF) collected from the cerebellomedullary cistern after MRI was found to be within normal limits. A PCR-based panel evaluating CSF and whole blood for canine distemper virus (species Canine morbillivirus), West Nile virus, Borrelia burgdorferi, Neospora caninum and N. hughesi, Toxoplasma gondii, Anaplasma phagocytophilum, Ehrlichia canis, and Rickettsia spp. as well as serology for N. caninum were negative. Creatine kinase (CK) was markedly elevated by the following day (Table 1). Urinalysis revealed amber-colored urine with marked hematuria (250 × 10⁹ RBC/L) and proteinuria (1.0 g/L). The dog was treated with intravenous fluid therapy, and within 3 d the dog’s neurologic conditions had improved significantly, with only minimal residual ataxia, and the hematuria had grossly resolved allowing discharge of the dog. Two wk after ingestion, the owner reported that the dog had recovered fully and resumed normal activity. Laboratory tests did not reveal common toxic and infectious etiologies or a CNS mass as the cause of this dog’s acute neurologic crisis, suggesting that the likely cause was consumption of the hydrogel material in the commercial dog pad. The limited availability of information about the potential toxicity of polyacrylic acid hydrogels in this commercial product prompted our follow-up rat study, which tested the hypothesis that oral exposure to this polyacrylate-based hydrogel can elicit toxicity in mammals.

Materials and methods

Postmortem examination and neuropathology

Rats (n = 6; Table 2) were prepared for neuropathology assessment using methods published previously. ¹ Rats were anesthetized with sodium pentobarbital (150 mg/kg given by intraperitoneal injection) and perfused with ~0.15 L of a heparinized saline flush solution followed by fixation with ~0.25 L of 4% methanol-free formaldehyde (reconstituted from paraformaldehyde powder) in 0.05 M sodium phosphate buffer (pH 7.4). Solutions were maintained at 4°C until immediately prior to perfusion. The brain, spinal cord, and one sciatic nerve with its main branches were exposed, examined grossly, and immersed in fresh fixative at 4°C for at least 24 h prior to histologic processing for neuropathology evaluations. Dorsal root ganglia; spinal nerve root fibers; brain (olfactory bulbs, cerebral cortex [associative, motor, and sensory areas], striatum, hippocampus, thalamus, hypothalamus, cerebellum, pons, and medulla oblongata); spinal cord (2 blocks from the lumbar intumescence); and sciatic nerve (adjacent to the sacrum and just proximal to the stifle) from one hindlimb were embedded in paraffin and sectioned serially at 5–6 μm to permit examination by bright-field light microscopy. Sections were stained with hematoxylin and eosin (to permit an assessment of general architecture) or anti-glial fibrillary acidic protein (to reveal any early glial reactions).

OPEN IN VIEWER

Table 2.

Estimated exposure of individual rats consuming a pet pad hydrogel.

Rat ID	Estimated ingested dose (g/kg)	Onset to clinical signs*	Initial clinical signs	Outcome
A	19.2	7 h	Limp body or reduced reaction to handling, delayed PLR, change in gait/walking on toes, hunched body posture, ataxia, decreased activity in the FOB open field.	Euthanized on day 3†
B	18.6	7 h	Decreased activity in the FOB open field, no approach-response reaction.	Euthanized on day 3†
C	11.9	6 h	Muscle tremors (head), change in gait or walking on toes, hunched body posture, decreased activity in the FOB open field.	Euthanized on day 3†
D	11.8	4 h	Decreased muscle tone, limp body/reduced reaction to handling, dilated pupils with delayed PLR, change in gait or walking on toes, ataxia, CNS depression, decreased activity in the FOB open field.	Found dead at 28 h (alive at 26 h)
E	10.9	6 h	Decreased activity in the FOB open field.	Euthanized on day 3†
F	7.6	6 h	Limp body or reduced reaction to handling, change in gait or walking on toes, CNS depression, hunched body posture, decreased activity in the FOB open field.	Euthanized at 28 h†
G	5.9	3 d	CNS depression, ataxia, splayed gait, distended abdomen.	Euthanized on day 3†
H	5.5	26 h	Decreased muscle tone, limp body or reduced reaction to handling CNS depression, change in gait or walking on toes and splayed, hunched body posture, ataxia, decreased activity in the FOB open field.	Euthanized at 28 h
I	2.6	32 h	Decreased muscle tone, limp body or reduced reaction to handling, CNS depression, change in gait or walking on toes, hunched body posture, ataxia, decreased activity in the FOB open field, no approach-response reaction.	Found dead on day 3

CNS = central nervous system; FOB = functional observational battery; PLR = pupillary light reflex.

*

Time when clinical signs were first noted in the FOB or cage-side observation periods.

†

Perfusion and neuropathology performed.

Results

All pre-dosing FOB evaluations (on day −5) were within normal limits. All rats licked the peanut butter off the hydrogel immediately upon presentation. Afterwards, all rats consumed portions of the hydrogel starting at 2 h after the start of the feeding trial. The estimated mean amount of hydrogel material ingested was 10.4 ± 5.7 g/kg (range: 2.6–19.2 g/kg; Table 2). The most common neurologic abnormalities observed during day 0 FOB included decreased activity in the FOB open field, changes in posture and reactivity to handling, and change in gait or walking on toes (Table 2). Two rats that consumed 5.9–18.6 g/kg of the hydrogel had no abnormal findings during day +1 FOB. All other rats demonstrated one or more abnormalities during day +1 FOB. Hydrogel ingestion was also associated with decreased motor activity, reduced numbers of rears, decreased body temperature, and decreased body weight on day +1 (Table 3).

OPEN IN VIEWER

Table 3.

Quantitative functional data in hydrogel-exposed rats.

Parameter	Unit	Pre-exposure*	Post-exposure (day +1)	Paired t-test result
Body weight	g	209 ± 12 (−5)	205 ± 12	t(7) = −6.06, p = 0.0003
Body temperature	°C	37.6 ± 0.4 (−5)	34.4 ± 3.7	i(7) = −2.42, p = 0.046
Forelimb grip	kg	0.48 ± 0.09 (−5)	0.38 ± 0.11	t(6) = −1.44, p = 0.201
Hindlimb grip	kg	0.23 ± 0.07 (−5)	0.28 ± 0.04	t(6) = 1.36, p = 0.221
Hindlimb splay	mm	8.0 ± 0.5 (−5)	4.9 ± 0.6	t(4) = −4.15, p = 0.014
Distance traveled	m	45.3 ± 5.2 (−6)	9.5 ± 12.0	t(8) = 8.78, p < 0.001
No. of rears		54 ± 15 (−6)	7 ± 10	t(8) = 9.09, p < 0.001

Values are given as mean ± standard deviation.

*

Values in parentheses are days before the exposure was initiated.

Histologic structures within examined sections of the nervous system for all 6 rats that survived to day +3 or 1 of 2 rats that were euthanized on day +2 (Table 2) were all within normal limits.

The ¹H-NMR spectrum of the hydrated hydrogel is consistent with the polyacrylic acid reference sample but also shows peaks in the 6–8 ppm range that do not appear in the polyacrylic acid reference sample (Fig. 1). These peaks are also seen in the ¹H-NMR spectrum of the D₂O extract. The IR spectrum of the lyophilized gel also suggests the presence of alkenes and/or amino groups (Fig. 2).

OPEN IN VIEWER

Figure 1.

¹H-NMR spectra of the deuterium oxide extract of the hydrated hydrogel from the commercial dog pad (top), polyacrylic acid reference standard (middle), and the hydrated hydrogel from the commercial dog pad (bottom).

OPEN IN VIEWER

Figure 2.

Fourier-transform infrared spectra of polyacrylic acid salt reference standard (top) and lyophilized gel sample taken from the commercial pet pad (bottom).

Discussion

An adult, male Australian Shepherd mix that was presented to NCSU-VH had developed vomiting, altered mentation, ataxia, intention tremors, and tachypnea within several hours following the ingestion of ≤15.7 g/kg of the hydrogel filling a commercial pet pad. Neurologic signs consistent with suspected acute forebrain and cerebellar neurotoxicity persisted for at least 96 h before the dog began to recover slowly. Data from the ASPCA Animal Poison Control Center indicate that a mixture of gastrointestinal (e.g., vomiting, diarrhea) and neurologic (e.g., ataxia, tremors, hyperesthesia) signs are often observed in dogs consuming hydrogels (Dr. T. Wismer, pers. com., 2017). When considered together, these findings suggest that polyacrylic acid hydrogel ingestion may induce clinically significant gastrointestinal irritation and neurotoxicity in dogs.

Evidence of acute neurotoxicity was also observed in rats that ingested the pet pad hydrogel. The onset of neurologic signs in rats appeared to be dose- and time-dependent. Rats ingesting >7 g/kg of the hydrogel developed behavioral changes or abnormal neurologic signs within 6–12 h after ingestion, whereas rats ingesting smaller amounts of hydrogel (2.6–5.9 g/kg) had no abnormal clinical or neurologic findings 12 h after ingestion. Ingestion of hydrogel by rats was often associated with one or more abnormalities on the neurotoxicity test battery used. The endpoints used in our study (FOB, motor activity, and neuropathology) are commonly incorporated in rodent neurotoxicity test batteries.^8,15,17 Abnormal responses seen in rats during the post-exposure FOBs predominantly involved changes in neuromuscular function and coordination. Despite finding reduced muscle tone during the post-exposure FOBs, changes in grip strength were not observed during the study. This result is somewhat surprising given that hindlimb grip strength measures digital flexor muscle responses initiated by the grasp reflex and can be correlated with muscle tone.^12,18 Decreased spontaneous motor activity and reduced rearing behaviors were also seen in the hydrogel-exposed rats. Spontaneous locomotor activity is an apical test of neuronal function and is a sensitive measure of nervous system effects.^15,19 Although decreased motor activity is often associated with chemicals that affect the nervous system, changes in motor activity cannot be attributed to specific neuronal substrates. ¹⁵ Previous investigators have reported that poor clinical condition is not always associated with decreased motor activity in rodents, suggesting that the hydrogel may have direct effects on the nervous system. ¹⁰ Ataxia, splayed gait, slowed surface righting, and slowed pupillary light reflexes were also observed in some affected rats, which would further support the conclusion that the hydrogel is neurotoxic. None of the rats that ingested the hydrogel developed anatomic lesions in brain indicative of acute neurotoxicity (e.g., neuronal necrosis, increased glial reactivity). This finding is consistent with the lack of structural brain MRI changes seen in the dog that ingested hydrogel. Additional studies are needed to determine whether changes in brain neurochemistry could account for the observed neurologic effects.

Ingestion of the hydrogel may have additional effects on other organs or systems. Some of the changes in posture (hunched appearance) and gait (walking on tiptoes) seen in hydrogel-exposed rats during the post-exposure FOBs may be indicative of abdominal discomfort and pain in rats. ¹¹ If pain was present, then the most likely cause is marked abdominal distension resulting from expansion of the hydrogel in the rat’s stomach as a result of water absorption. Expanded hydrogel resulting in gastric dilation was observed in several animals that died or were euthanized within 48 h of hydrogel ingestion. This gastric dilation effect may be more prominent in rats given that, unlike dogs, they cannot vomit. ⁶ Gastric impaction from the hydrogel may have contributed directly to some clinical signs and the death of exposed rats. Fatal gastrointestinal impaction has previously been reported in birds that consumed a polyacrylamide-based hydrogel. ¹³

One of the limitations associated with our case is identification of the neurotoxic chemical found in the pet pad hydrogel. Based on the MSDS provided by the manufacturer, the hydrogel contains a 30:70 mixture of polyacrylic acid and water. Chemical analyses (nuclear magnetic resonance and IR spectrometry) of the gel from our case confirmed the presence of polyacrylic acid along with other unidentified compounds. For example, the presence of the ¹H-NMR peaks seen at 6–8 ppm suggests that leachable short carbon chain (3–8 carbons) molecules in the gels contain carbon–carbon double bonds, oxygen, and possibly nitrogen. Although the oxygen seen in the NMR can be associated with some polyacrylic acid chains present in the extracts, the short carbon–carbon chains with double bonds and amino groups are likely to be external contaminants. Although the shift of the carbonyl peak in the IR spectra from 1.539 (carboxylic acid salt) to 1.645 m^-1 is consistent with gel formation, ²⁰ the additional peaks at ~1.6 m^-1 indicate the presence of alkene and/or amine groups, and the peaks at ~3.3 m^-1 are indicative of primary amino groups. The Chemical Abstract Services (CAS) number assigned to the product (CAS 9003-04-7) is used for a broad family of structurally related polymers composed of the sodium salt of polyacrylic acid. In general, CAS registers polymers based on their structural repeating unit (e.g., in our case, the repeating unit is [C₃H₄O₂]_x•xNa). The exact composition and physicochemical properties associated with these polymers can vary based on polymer length, crosslinking, and other factors. ²¹ Thus, conclusions regarding hydrogel toxicity drawn from our study are applicable to the formulation found in the commercial pet pad product that was tested in our laboratory and may not be applicable generally to other products that contain polyacrylic acid–based hydrogels. Further work will be required to evaluate whether the clinical picture produced by this hydrogel is specific to its chemical composition or may be ascribed more generally to its physical properties as a hydrogel.