The different AC formulations tested herein similarly decreased concentrations of the majority of assayed toxicant in simulated gastrointestinal fluid with both acidic and neutral pH. Within 30 min of addition of all AC products, more than 96% of naproxen and 98% of ivermectin was adsorbed in both environments. Naproxen ingestion above 25 mg/kg is expected to result in renal toxicity in dogs, and we used a total of 880 mg in the flask (9). The tested dose of ivermectin was 114 mg, and severe neurologic symptoms are expected to manifest at doses between 1–2.5 mg/kg in dogs with a functional MDR gene (and result in more severe toxicity in a dog with non-functional MDR gene) (10). While the tested dosage of 15 g of AC is consistent with a dose that might be administered to a 10 kg. dog, almost all toxicant was adsorbed in the test flasks. While RR and pAC contain only AC, Tox is a proprietary mixture of kaolin and AC, supplied either as a liquid with 10.4% AC and 6.25% kaolin, or as granules containing 47.5% AC, 10% kaolin, and 20% sorbitol. Based on the results shown here, we suspect that administration of AC products after oral consumption of naproxen or ivermectin would adsorb the majority of remaining drug in the stomach and intestine.

Delta-9 THC was also adsorbed to an equivalent degree by all AC types in both acidic and neutral pH environments. The LD₅₀ for THC in dogs has not been determined, and the amount of active ingredient in many products can be variable and unpredictable (11). We chose a baked product with a stated amount of Δ-9 THC, but the chances that the compound was evenly distributed throughout the brownie are not known. This is important because equivalent weight pieces were divided in thirds for the assays, but variation in Δ-9 THC content for each specific piece was not known.

Bromethalin was less well adsorbed over the study time period compared to other toxicants, but all AC products adsorbed the bromethalin equivalently in both acidic and neutral pH environments. The persistent bromethalin detected in the treated flasks may have been a result of continued elution from the bait blocks that were used as a source, as they were not removed prior to introduction of the AC. We used 2.8 mg of bromethalin in each assay which was the amount contained in a single block of rodenticide [the LD₅₀ of bromethalin is 2.38–5.6 mg/kg in dogs (12)], and the blocks were not crushed prior to addition (although they visually appeared to dissociate), which may have contributed to the prolonged elution of bromethalin. An additional possibility for the lack of significant decrease in bromethalin concentration can also be due to binding and subsequent dissociation of the toxicant from the AC; we did not study the kinetics of binding of any toxicant to AC but this may also have contributed to the spikes in bromethalin concentration seen at the 60 and 90 min timepoints for some AC types.

Differences in the rate of adsorption between AC types were seen for theobromine and caffeine derived from baker’s chocolate. The flasks with RR demonstrated a slower rate of decrease in concentration compared to that of Tox or pAC, specifically slower than pAC for theobromine in the acidic environment and Tox for theobromine in the neutral pH. Caffeine in the acid pH was adsorbed similarly by all activated charcoal products after 4 h, but the rate of caffeine concentration decrease was more rapid in pAC compared to RR. In the neutral pH the concentration of caffeine decreased in a similar manner, but any differences in rate between the AC products did not reach statistical significance. Theobromine and caffeine as tested here were components of baker’s chocolate, in this case administered as 56.7 g of chocolate (13). Given that chocolate also contains cocoa butter and other large molecules, the real and apparent slower adsorption in the RR group may also represent steric hindrance or coating of the granules by the fats in the compound. It will be beneficial to make further observations of adsorption efficacy in the presence of possible interfering substances, and it is probably wise to avoid administering RR to dogs using fatty admixtures (e.g., peanut butter) as opposed to mixing with water or possibly small amounts of food, to avoid any possible obstacle for access to the interior pore structure. The molecular weights of caffeine and theobromine are smaller than other compounds that were adsorbed well by RR (e.g., ivermectin, see Table 7), so we do not believe that the size of the molecules played a role in the slower adsorption.

Due to low concentrations, the adsorption of roquefortine was not able to be compared between AC types, and the only data is based on extrapolation of signal that was below the reporting limit of the laboratory that ran the assay. While there appeared to be decrease in concentration in all assays of roquefortine (from detectable to undetectable), caution is warranted in extrapolating this information to clinical cases. The toxic dose of tremorgenic mycotoxins such as roquefortine in dogs has not been identified (14), and we used 50 mg of aged Roquefort cheese in each trial. While the source was authentic Roquefort cheese, additional trials with larger doses of pure toxicant are indicated, both as an improved assay and to verify equal amounts of toxicant to each assay (assuming that the roquefortine was not evenly distributed throughout the cheese).

In this study, the source toxicant compounds were chosen to mimic an actual animal ingestion, instead of studying adsorption of isolated toxicants. Although this was a more realistic usage of the products, the impact of excipient products (e.g., the cocoa butter in chocolate) that may have altered the adsorption kinetics could not be accounted for. Similarly, we suspect that the continued increase in concentration of toxicant in some assays represented continued elution of toxicant from the source compound, rather than a representation of saturation of charcoal binding sites on the AC products, but this must be considered as another possibility. Likewise, depending on the compounding and the specific assay, we cannot rule out interference from the components of the toxicants on the assays. Specifically, some components of the ivermectin paste were detected along with the ivermectin on the HPLC output from the assays, and a more specific assay (e.g., GC–MS) may be less likely to demonstrate interference. Future testing to determine actual adsorptive efficacy should be focused on comparisons between known quantities of pure compounds.

All of the tested compounds had molecular weights (MW) between 67.07 g/mol and 875.1 g/mol. Though there is no published report on the exact size range of molecules activated charcoal binds, previous studies have utilized compounds with molecular weights between 100–800 g/mol (15). The binding of compounds to AC is thought to be through intermolecular or van der Walls forces. Van der Walls forces describe attraction of compounds without any change in the structure of the compound and can occur between both uncharged and charged molecules, with charged molecules potentially able to form stronger bonds with the uncharged AC. The impact of assay pH on the ionization of the compounds may thus have impacted the tightness of binding to the AC molecules and ability to saturate available binding sites, which could have impacted the apparent adsorption as measured here. Additionally, the overall solubility of some compounds in aqueous solution may have hindered dissociation of individual compounds, although the acid and neutral fluids did contain bile salts and other emulsifiers as might be found in stomach or intestinal secretions, and all toxicants were able to be detected on the assays, with only roquefortine being found at very low concentrations. The relative lipophilicity (as reflected in the octanol–water partition coefficient, logP) of the tested compounds is also listed in Table 7. Although the lipophilic compounds (logP>3) did exhibit binding to all AC formulations, as expected, some non-lipophilic compounds (e.g., caffeine, theobromine) also showed significant adsorption, which may be a factor of other molecular characteristics (e.g., molecular geometry).

The assayed alcohols, xylitol and ethylene glycol, were not expected to show evidence of adsorption to charcoal. The increasing xylitol concentration at the 4 h points was hypothesized to be due to continued elution from the gum, as it dissolved more slowly than other toxicants. Although it was expected that the ethylene glycol concentrations would not decrease, the RR and AC products did show decreased concentrations after longer incubation in both the acidic and neutral pH environments. Similarly, the tartaric acid concentrations, which increased in both pH environments exposed to RR and AC, showed a decrease to about 30% of baseline in the Tox group with acidic pH, and to 60% of baseline in the neutral pH environment. These results, as well as the decrease in the assays with some of the alcohols, may be due to the presence of kaolin in the Tox mixture, or due to the ionization of tartaric acid in the neutral pH. Tartaric acid was chosen for study due to its putative implication in the renal toxicity caused by ingestion of grapes, raisins, or tamarind (16). Kaolin is a clay composed of aluminum silicates that is useful for binding of compounds in the gastrointestinal tract but does not create chemical bonds in the same manner as AC, rather resulting in bulk sequestration of toxins and other compounds. Because of the relative lack of specificity of kaolin, it may have an enhanced ability to bind alcohols, and may work synergistically with AC.

Ready Rescue has a similar surface area as Tox or AC but a higher density, with part of the surface area contained within the spherical granules. The exact surface area of the Tox and pAC that were tested in this investigation is not known, and so the actual adsorptive surface area per weight of AC may have been different. We also encountered a backorder of Tox, which forced us to change products from the granules to the solution midway through this experiment. Notably, the granules contain 10% Kaolin, 20% Sorbitol and 22.5% unidentified wetting and dispersing agents in addition to 47.5% AC, while the tested suspension contained only 10.4% AC and 6.25% kaolin in an aqueous base. While the application was based only on equivalent amounts of AC, we cannot predict the impact of the other ingredients on adsorption of toxins between the two Tox products.

Other limitations include the fact that due to logistics and funding, only single adsorption trials were performed for each toxicant and timepoint. Optimally these would have been performed in replicates. Based on the results seen here, future studies can focus primarily on the 30 and 60 min timepoints, especially because the toxicant will likely be absorbed or moved out of the stomach at the 4 h time point under circumstances of an in vivo unplanned ingestion. Additionally, to mimic an actual pet ingestion scenario, commercially available products, in various substrates or vehicles were tested. To obtain specific information about the exact ability of the different AC products to adsorb the tested toxicants, pure compound should be used. Using a set amount of pure compound will also simplify data analysis, as it is suspected that the continued increase in concentration of some toxicants was related to continued elution of the compound from the substrate. Due to the lack of pure toxicant used here, we also cannot predict the impact of the specific preparations on the analysis. Although samples were grossly purified through centrifugation prior to analysis, interfering compounds from the toxicant vehicles may have caused some degree of interference in the final assays.

In addition, the impact of admixture of any AC with food (as is commonly used to encourage ingestion of AC in pets) was not evaluated. Depending on the composition of the dog food, specifically of lipid soluble ingredients, it is plausible that the addition of lipid soluble components might impact adsorption of the actual toxin. One product (RR) has guidance that admixture with dog food does not impact adsorption while admixture using peanut butter may decrease toxicant adsorption. Whether this is true across AC types or if it is a function of the spherical shape of RR is not clear, and requires further study. Without further study, the investigators can only recommend administering an AC product admixed with water, or without admixture.

The in vitro data presented here suggests that RR has similar adsorptive capacity to Tox and pAC for common canine toxicants, but these in vitro results may not reflect in vivo effects, and further in-vivo testing would be necessary to confirm clinical adsorptive efficacy.