Materials and Methods

Horses—Four healthy adult neutered horses (2 geldings and 2 ovariectomized mares) from the Oregon State University research herd aged 4, 10, 15, and 19 years of age and weighing (mean ± SD) 531.8 ± 21.3 kg were available for use in the study. Physical examinations and preliminary blood analyses including a serum biochemical profile and venous blood gas analysis were performed. All horses were considered healthy at the time of the study. The horses were brought into the Oregon State University Large Animal Clinic for acclimation 2 days before the trial and were housed in individual stalls in a climate-controlled barn. Diet included grass hay fed twice daily and ad libitum water for the duration of the study. The horses were housed and cared for in an accredited facility, and the study was approved according to the principles outlined by the Oregon State University Institutional Animal Care and Use Committee.

Study design—One day prior to the start of each treatment, an IV catheter was placed in the jugular vein of each horse to facilitate blood sample collection. Three treatments were administered to each horse in random order in a crossover design with a 1-week washout period between treatments. The 3 treatments were IV administration of metformin (6 g), PO administration of metformin (6 g) after feed withholding for 12 hours, and PO administration of metformin (6 g) with no feed withholding. For IV administration, 6 g of metformin^a was dissolved in 15 mL of saline (0.9% NaCl) solution and passed through a 0.2-μm filter.^7,b In a previous study⁷ that used the same technique, the solution for IV use was analyzed by use of HPLC (before and after filtration) and found to contain 98% of the stated dose. Intravenous treatments were administered in 500 mL of saline solution over 5 minutes via the jugular catheter within 1 hour of compounding. The catheter was flushed thoroughly with heparinized saline solution after metformin administration. Oral treatments were administered by crushing metformin tablets^c and mixing with water and administering by mouth with a syringe within 1 hour after compounding.

Blood samples were collected at times 0, 0.5, 1, 2, 4, 6, 8, 10, 12, 14, and 24 hours after treatment. For the IV treatment, samples were also taken every 15 minutes for the first 4 hours. To preserve stability, plasma samples were obtained by centrifugation of blood within 1 hour after collection and immediately stored at −80°C for ≤ 3 weeks until shipment to a commercial laboratory^d for analysis of metformin concentrations via HPLC. Additional blood samples were collected for measurement of blood lactate concentrations at 0, 4, 8, 12, and 24 hours after IV metformin administration and placed in lithium heparin–containing tubes. Intravenous jugular catheters were removed at the end of each testing period.

Assay—Quantitative analyses for metformin in plasma were performed by a commercial laboratory^d with an HPLC procedure following derivatization with p-nitrobenzoyl chloride.¹⁹ In summary, an internal standard (buformin, an antidiabetic drug chemically related to metformin) was added to a 0.5-mL aliquot of specimen. Following the addition of sodium hydroxide and approximately 0.6 g of sodium chloride, the sample was mixed well to dissolve the salt. One milliliter of p-nitrobenzoyl chloride reagent (4 g of para-nitrobenzoyl chloride in 100 mL of acetonitrile) was added, the tube was capped, and the solution was mixed for 15 minutes and extracted with a mixed solvent of 30% ethyl acetate in n-butyl chloride. The solvent was removed and back extracted in 1N hydrochloric acid, which was made basic with sodium hydroxide, and again extracted in the mixed solvent. The solvent was evaporated to dryness, and the residue was reconstituted with mobile phase. Extracts were analyzed via HPLC with UV detection at 280 nm. To validate the assay in equine plasma, 6 concentrations of calibrators ranging from 0.05 to 50 μg/mL were included with each analytic batch, and 2 concentrations of control samples were included with each analytic batch. The between-run precision was 13.6% and 8.48% at 1.0 and 2.5 μg/mL, respectively. The lower limit of quantitation was 0.05 μg/mL, and the upper limit of quantitation was 50 μg/mL. The interassay accuracy was 95% at 1.0 μg/mL and 96% at 2.5 μg/mL. Intra-assay precision was 8.3%, 2.3%, and 5.2% at 0.05, 1.0, and 2.5 μg/mL, respectively. Intra-assay accuracy was 97%, 103%, and 94% at 0.05, 1.0, and 2.5 μg/mL, respectively.

Data analysis—Metformin concentration–time data following IV and PO administrations were analyzed by use of noncompartmental analysis and estimated by use of a computer program.^e The C_max and time of maximal concentration were determined via observation. The AUC was calculated by use of the linear trapezoidal method. Volume of distribution, elimination rate constant, mean residence times, and clearance were estimated by use of noncompartmental analysis with IV input and first-order output from the plasma concentration–time data collected during IV treatment from 15 to 120 minutes. To determine the bioavailability of metformin, the absorption of the drug after PO administration was compared with the concentration-time relationship (AUC) in the bloodstream after it was injected IV. Bioavailability of orally administered metformin was reported as a fraction of the concentration reached after IV administration (F = AUC_oral/AUC_IV). Bioavailability was determined with and without feed restriction. Values are reported as mean ± SD.

Results

Results of the pharmacokinetic analysis (Table 1) revealed that, following IV administration, metformin was rapidly eliminated from the plasma (Figure 2). The concentration half-life was quantifiable for > 5 half-lives. The mean distribution half-life of metformin following IV administration was 24.9 ± 0.4 minutes with a mean residence time of 27.5 ± 2.1 minutes. The mean AUC was 20.9 ± 2.0 h·μg/mL for IV administration, whereas PO administration resulted in AUCs of 1.6 ± 0.4 h·μg/mL in unfed horses and 0.8 ± 0.2 h·μg/mL in fed horses. The AUCs determined for each horse were used to estimate the bioavailability, and the mean bioavailability of metformin was approximately 7.1 ± 1.5% in unfed horses and 3.9 ± 1.0% without feed withholding. This estimate was based on the plasma data collected from fed and unfed horses, compared with IV infusion for those samples with quantifiable amounts of metformin. Quantitative data were obtained for samples collected up to 8 hours following administration. The maximal concentration of metformin achieved with the 6-g dose following PO administration was 0.4 ± 0.1 μg/mL in unfed horses. The C_max was reduced to 0.3 ± 0.04 μg/mL in fed horses; however, this was not significantly different from values in unfed horses.

Mean plasma concentrations of metformin after IV and PO administration of 6 g to 4 adult horses. Oral administration was performed when horses were fed and also when feed was withheld for 12 hours. lg = Log.

Citation: American Journal of Veterinary Research 70, 5; 10.2460/ajvr.70.5.665

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Mean plasma concentrations of metformin after IV and PO administration of 6 g to 4 adult horses. Oral administration was performed when horses were fed and also when feed was withheld for 12 hours. lg = Log.

Citation: American Journal of Veterinary Research 70, 5; 10.2460/ajvr.70.5.665

• Download Figure
• Download figure as PowerPoint slide

Mean plasma concentrations of metformin after IV and PO administration of 6 g to 4 adult horses. Oral administration was performed when horses were fed and also when feed was withheld for 12 hours. lg = Log.

Citation: American Journal of Veterinary Research 70, 5; 10.2460/ajvr.70.5.665

• Download Figure
• Download figure as PowerPoint slide

Adverse effects of treatment—Mild signs of abdominal discomfort (pawing) and loose feces, characterized as an increased water content with loss of fecal ball structure with no change in fecal consistency, were observed in 2 horses within 4 hours after IV administration of metformin. Signs resolved within 15 minutes after walking the horses. No clinically evident adverse effects were observed after PO administration of the drug. Blood lactate activity remained within reference range (0.5 ± 0.1 mmol/L) after IV administration in all horses studied.

Discussion

Results of the present study revealed marked differences in the pharmacokinetics of metformin in horses, compared with humans. In humans, metformin has a half-life of 6.2 hours as judged on the basis of a 3-compartment open model with initial elimination half-life estimated to be 1.7 to 3.0 hours and a terminal elimination of 9 to 17 hours.⁸ In the present study, metformin was rapidly eliminated and no longer detectable in plasma after 4 hours. Following IV administration in humans, the drug is completely excreted unchanged in urine with a renal clearance of (mean ± SD) 454 ± 47 mL/min.⁸ Our estimates suggested that the clearance of metformin in horses is > 10 times that in humans and may, in part, explain the low plasma concentrations following PO administration.

Results of a human study²⁰ indicate an absolute oral bioavailability of 40% to 60%, with gastrointestinal absorption complete within 6 hours of ingestion. In rats and humans, the rate of absorption is slower than the rate of renal elimination, which results in what has been described as a flip-flop pharmacokinetic model.^8,21 Furthermore, an inverse relationship between the dose ingested and the relative absorption has been observed as doses are increased,^20,21 suggesting the involvement of an active, saturable absorption process that we hypothesize may be lacking in horses. On the basis of our results, it is possible that the rate of absorption and clearance of metformin may be higher in horses than in humans, which may in part account for the drug's low bioavailability. Low fractional bioavailability, as measured in the study reported here, can be caused by either low absorption or high excretion. Additional studies, including analysis of urinary excretion data, are needed to determine which of these is responsible for the low bioavailability of metformin in horses. Despite the rapid clearance from plasma, the maximal concentrations reached were close to the reported therapeutic range for humans of 0.5 to 1 μg/mL.²⁰ Our data suggested that dosages and administration intervals used in earlier equine studies were insufficient to achieve plasma concentrations of drug comparable to the therapeutic range achieved in humans. More information regarding metformin pharmacodynamics and pharmacokinetics in horses may be obtained by use of a larger sample size.

Low bioavailability only partially explains the lack of success with administration of metformin in horses. Metformin is reported to have a substantial first-pass pharmacodynamic effect that occurs in the liver and gastrointestinal wall in rats, and PO administration of metformin provides greater clinical response than IV administration in humans.^22,23 These pharmacodynamic and pharmacokinetic effects are hypothesized to be linked to the therapeutic effects of treatment because stronger, sustained glucose-lowering responses are observed when metformin is administered orally, compared with the equivalent dose administered IV.²² On the basis of the results of the study reported here, we hypothesize that the lack of clinical success of treatment with metformin in horses may be attributable to a combination of low bioavailability and increased rate of elimination. Although the horses used in this study had a wide age range, we did not detect a significant difference in bioavailability in young versus old horses. Therefore, it is apparent that a study designed to measure efficacy of metformin in horses may require larger doses and more frequent administration or perhaps a sustained-release formulation.